Author: George Weiner
Date: 2015-02-24 10:07:41

Dr. Laura Rogers, a post-doc in my research lab, likes to start her presentations during our weekly research laboratory meetings with a quote. One she used a few months back was from Thomas Huxley, a renowned British biologist from the 1800s, who said, “The great tragedy of science – the slaying of a beautiful hypothesis by an ugly fact.” It was a very appropriate quote for Laura to use given the results we were discussing that day.

As Huxley’s quote illustrates, Laura wasn’t the first scientist to see a beautiful hypothesis slain by an ugly fact, and she will not be the last.

Here is the story. Treat vWe know that everybody’s immune system is unique. There is strong evidence that individual differences in the immune system can impact many aspects of the immune response. My research group has been interested for some time in understanding why the anti-lymphoma antibody known as rituximab works better for some patients than others. Figuring this out would help us understand how we can use the immune system more effectively to treat cancer.

We therefore took our large collection of samples from lymphoma patients who had been treated with rituximab to see if individual differences in their immune systems might explain why rituximab did, or did not, work well for each individual. This was very much an exploratory study since we did not have preconceived notions about which particular part of the immune system might be most important. In scientific terms, we considered this “hypothesis generating” research.

The results were promising. They suggested that patients with higher levels of a protein known as CFHR1 have a better response to rituximab. Other researchers have shown that CFHR1 levels are associated with a variety of immune system diseases, which spoke to the potential importance of this protein. Our finding led to the hypothesis that was at the center of Laura’s research project, namely that “CFHR1 enhances the response of lymphoma patients to rituximab.”

We set off to explore this hypothesis in two ways. First, we looked in the laboratory to see if CFHR1 enhances how well rituximab works in the test tube. Second, we looked to confirm our original results to determine whether higher levels of CFHR1 correlate with a better clinical response to rituximab using samples from a totally different group of patients. Laura got to work on both parts of the project.

Because no one had produced pure CFHR1 that we could use for our experiments, Laura produced it herself. At the same time, we worked with a pharmaceutical company to obtain samples from a clinical trial. This company had samples from patients treated with rituximab in Europe and we used these samples to confirm that high levels of CFHR1 correlate with a better response to rituximab. To do this, Laura measured CFHR1 levels on blood samples from all of these patients.

You can probably figure out the results at this point - so much for our beautiful hypothesis. First, the test tube results failed to demonstrate that CFHR1 enhances the anti-lymphoma effects of rituximab. Second, the attempt to confirm our observation that CFHR1 levels correlated with response to rituximab fell flat on its face. There was no correlation at all between CFHR1 levels and response to rituximab in this second set of samples. The simplest explanation for our findings is that the correlation we observed in our original study between CFHR1 and response to rituximab was simply chance.

As Huxley said, our beautiful hypothesis had been slain by ugly facts. Does this tale imply that this research project was misconceived? Not at all. Indeed, science is all about generating and testing hypotheses. If all hypotheses proved to be true, there would be no purpose behind testing them in the first place.

Laura took this very well. She is writing up her experiments for publication (even though the results failed to support our hypothesis) and has now moved on to another project. She has come up with another beautiful hypothesis. Only by doing the research will we know if our new hypothesis will survive the merciless, but absolutely critical, judgment of the facts.