In summer 2020, Michelle and I decided to fund research into cannabis and ovarian cancer. There were no existing studies that accounted for both: the complexity of cannabis and the complexity of ovarian cancer. Michelle made a connection with Prof. Hinanit Koltai of the Volcani Institute in Israel, and a collaboration was born.
Today, the first fruits of that collaboration have been published! You can read the full paper here.
Michelle desperately wanted (and deserved!) to see this research in print, so this moment is bittersweet. Indeed, for a long time we thought it could be published by summer 2021; we never imagined that publication would take more than another year. I know that Michelle would be thrilled; I so wish she were still here to celebrate.
So what does the paper say? There is a tremendous amount of information packed into it, as Hinanit and her team demonstrated cannabis's cancer-killing effects; identified likely modes of action; and examined how cannabis and existing chemotherapy could act synergistically. The presentation is aimed at specialist readers and so is quite terse; here and in subsequent posts I'll unpack some of the key results.
Part 1: what cannabis compounds kill ovarian cancer?
What are cell lines?
First, I need to make clear that none of the experiments (so far) are in people; rather, they used ovarian cancer cell lines--that is, cancer cells that were originally taken from a patient, long ago, and ever since have been maintained in lab cultures. A now-famous example of a cell line is that taken from Henrietta Lacks.
To match Michelle's cancer, we chose cell lines that were BRCA-negative. The majority of newer ovarian cancer drugs are focused on BRCA mutations, and are much less effective for BRCA-negative patients. We wanted to counteract that bias, and ensure that whatever we learned would be relevant to Michelle's disease. The main cell line analyzed in the paper was called HTB75 (also known as CAOV-3); many of the experiments were repeated on HTB161 (OVCAR-3).
Which cannabis strains kill ovarian cancer cells?
The first step was to evaluate the effects of whole-plant extracts--processing dried cannabis flowers in much the same way that cannabis oils or tinctures are made by commercial producers. These have a full spectrum of cannabinoids, terpenes and flavonoids, in whatever ratios were produced by the plants. The team tested five different strains that are used in Israeli medical treatments.
With each strain, the cannabis extract was applied in three different concentrations to batches of 10,000 cells, and after 2 days a measurement was made of the density of surviving cells. This measurement was compared to the value for a control treatment (no cannabis) to generate "% cell viability": the lower the number, the more cancer cells were killed by the treatment.
The bottom line: of the five strains tested, only one ("Dairy Queen") killed ovarian cancer cells.
Let me walk you through the figure that demonstrates this, as the paper is filled with similar figures. In the bar chart below, which is Figure 1A in the paper, the vertical axis is % cell viability and each bar is a different treatment.
The first bar ("control") represents cells treated with just the solvent used to dissolve the cannabis; because the solvent itself might kill some cells, this sets the 100% baseline.
The second bar is a "positive control:" a compound (in this case the chemotherapy agent Niraparib) that is known to kill ovarian cancer cells. This allows a confirmation that the experimental protocol can actually detect the effects of a cancer-killing agent.
The next three bars are different concentrations (low, medium, high) of extracts from the "Dairy Queen" strain (marked "DQ" on the figure).
The rest of the bars are the low, medium and high concentrations for the other four strains.
As you can see, mostof the bars are close to 100. The letters above the bars indicate groups of treatments that cannot be statistically distinguished: for example, all the bars marked "A" are indistinguishable from the control (that is, they do not kill cancer cells). Only two treatments led to reductions in cell viability: Niraparib (which is reassuring, as it is an existing treatment for ovarian cancer) and the highest concentration of Dairy Queen.
How does dose affect ovarian cancer cell viability?
The next graph (which I re-drew from figure 1B of the paper) shows how cell viability is affected by the dose (concentration) of Dairy Queen. Up to a concentration of 17.5 (which is midway between the medium and high levels of the previous graph), the cannabis had essentially no effect. However, above that level, increasing concentrations of the cannabis extract cause cell viability to drop off very quickly.
From this graph (which is often called a "dose-response curve"), the authors extract a value called IC50. This is the concentration at which 50% of the cells are killed--you can see it visually by looking at where the 50% line intersects the blue curve. Of course, our goal is to kill more than half the cancer cells! But IC50 is a standard value for summarizing these sorts of curves, and is easier to reliably measure than, say, the concentration required to kill 99% of the cells. In general, we can say that a compound with a lower IC50 can achieve the same cancer-killing benefit at a lower concentration (lower is better, especially when THC is involved!); it is used throughout the paper to compare various cannabinoid formulations.
What makes Dairy Queen different from the other strains tested? All the strains (including DQ) are relatively high in THC. Beyond that, I don't know a huge amount about what is and isn't in them. Dairy Queen is quite low in CBD, and relatively high in CBC (cannabichromene). CBC is not abundant in most strains, and, as we'll see, plays an important role in killing ovarian cancer cells. I also have a hunch (although there's no evidence for this in the paper) that CBD could interfere with THC's ability to act on these cancer cells (in contrast to other cancers where it may play an important anti-cancer role).
So does this mean that ovarian cancer patients should simply consume Dairy Queen? Well, maybe. There are several caveats. First, the chemical composition of a cannabis plant depends not just on its genetics, but also on its growing conditions. When Hinanit bought a new batch of Dairy Queen about a year after this experiment was run, the cannabinoid profile changed quite a bit different from from the first batch. Second, we don't yet know what kind of dose in people is equivalent to these doses in test tubes. And finally, we don't yet know if these results extend to ovarian cancers with BRCA mutations. So while consuming Dairy Queen probably wouldn't hurt, we don't know if it would help.
The rest of the paper is devoted to pulling apart Dairy Queen to identify the specific anti-cancer compounds; as well as mapping these effects onto cellular biochemistry.
What else is in the paper?
Phew, that was a lot of work to write! And that was just half a page of results. Now I understand why I was procrastinating on describing the preprint!
In future posts, I'll describe:
Which compounds (cannabinoids, terpenes, flavonoids) are driving the cancer-killing effect?
What is the interaction between cannabinoids and existing chemotherapy drugs?
How consistent are the effects across cell lines?
Are there detrimental effects on non-cancer cells?
What were the effects on cells from Michelle's tumor?
What biochemical processes seem to be involved in the anti-cancer activity of cannabinoids? (This is a large section of the paper, and may require several posts)
As I write these, I'll add links here.