It’s easy to focus on the downstream scientific effects of retractions. But sometimes they have financial implications, too.
Two weeks ago, we covered the retraction of a PNAS paper on a potential breast cancer treatment, one that would make tumors that didn’t respond to tamoxifen respond to the drug. We learned earlier this week from a Retraction Watch commenter that Wnt Research, a company based on the breast cancer finding and other work, was about to go public.
In fact, their initial public offering (IPO) happened today, and you can follow the price of their stock — listed on the Stockholm Stock Exchange as WNT — here. But what we learned when we looked into the IPO was that it was originally scheduled for late November, and was delayed because of the retraction.
Tommy Andersson, one of the researchers on the now-retracted paper and Wnt Research’s chief scientific officer, told Retraction Watch that the company had initially planned on going public on November 26. They had written a memorandum describing the company’s work to date, and its plans, and the public was given a chance to invest before shares hit the Stockholm exchange. That memorandum included a mention of the PNAS paper, as follows (translated from Swedish):
The research group has even shown that combining Wnt-5a or Foxy-5 will increase the expression of estrogen receptors in estrogen-negative breast cancer cells. It turned out that human breast cancer cells exposed to Foxy-5 regained tamoxifen sensitivity, which is clear from the increased apoptosis and reduced cell growth as a response to the endocrine treatment. Likewise, the research group found that administration of Foxy-5 in the body brought back the expression of estrogen receptors in a mouse model. Apart from the therapeutic potential of these observations, there is a potential to use the outcome of estrogen receptors as a biological marker, which should advance research.
The deadline for investment was October 27, and a number of people responded, allowing the company to continue its work. But on November 11, Andersson and his colleagues realized there were serious errors in the paper, and that it would need to be retracted. When Andersson called Wnt Research’s CEO, Bert Junno, on the 12th:
He rapidly called upon an extra Board meeting on the 15th of November (the same day my email of retraction was sent to the PNAS office). At this meeting we decided to make a press release on this matter and this went out on the 16th of November.
We again had an extra Board meeting on the 19th of November to discuss what other things that we needed to do. We agreed that the short press release must be accompanied by an addition to the already published memorandum, the reason being that you cannot change in the original memorandum. The CEO wrote this addition together with the people at the office of the small Stockholm Exchange “Aktietorget” where the company was to be listed. We also decided to ask for legal advice in how to handle the public that had already paid for shares in the company.
The board held another meeting on the 22nd, during which they approved the addition to the memorandum saying that the PNAS paper had been retracted. They also did something that can only be described as the right thing: They decided to write all of the approximately 275 people who had invested in the company by the October 27 deadline and offer their money back:
The reason for this was that the lawyer advised us to act according to good morals rather than to what we were required to do by law. His belief was that this would pay off in the long run.
The memorandum addition was published on November 25. And some people did ask for their money back:
The offer to retract their investments in the company resulted in a net loss of investments, but we still obtained enough investments to continue our work.
All of that back and forth also meant that the IPO was delayed by three weeks, until today.
We find this story wonderfully refreshing. Imagine, a company bending over backward to let investors — and potential investors — know about problems with its data.
That wasn’t what Arena Pharmaceuticals CEO Jack Leif did when he found out that the company’s weight loss drug lorcaserin was linked to cancer in rats. He decided that fact wasn’t “material,” or important enough to disclose to the public. Arena’s stock price got hammered for that decision, because the FDA ended up rejecting the drug over the potential cancer risk.
WNT lost funding in the short term, but it’s too early to tell what will happen to their stock price long-term. No matter what happens, they get a thumbs-up from Retraction Watch for transparency.
I know it’s been widely disseminated and assumed that the rat cancer issue was a big deal in the lorcaserin decision but I’m not sure that’s really the case. It certainly didn’t help, and I certainly agree it should have been disclosed earlier, but I think the key problem was the fear of heart valve problems and perhaps some other CV concerns, combined, of course, with efficacy coming in at the low end.
2016 correction figures 1,2 and 3.
Int J Cancer. 2003 Jan 20;103(3):344-51.
Wnt-5a and G-protein signaling are required for collagen-induced DDR1 receptor activation and normal mammary cell adhesion.
Dejmek J1, Dib K, Jönsson M, Andersson T.
Author information
1Experimental Pathology, Department of Laboratory Medicine, Lund University, Malmö University Hospital, Sweden.
2016 correction.
http://onlinelibrary.wiley.com/doi/10.1002/ijc.30258/full
On February the 25th 2016 the Journal of International Cancer notified us that “there may be irregularities in some of the figures in the above-cited paper published in the International Journal of Cancer”.
We fully agree with the Journal that the Western blots of the DDR1 loading controls in Figures 2A, 5A and 6A exhibit irregularities. The three senior authors of this paper regret that they did not detect these irregularities prior to submission of the manuscript.
To find the reason for these irregularities we studied the lab books from 2001. In doing so, we found the original blots used for the phospho-tyrosine blots outlined in Figures 2A and 6A. Next to the respective original phospho-tyrosine blot we found on the same page in the lab book three different exposures of DDR1 loading control blots for each of the phospho-tyrosine blots outlined in Figure 2A and Figure 6A. We cannot explain why these DDR1 loading control blots were not included in the original manuscript submitted to the journal. We have now used the correct DDR1 loading control blots in the corrected Figure 2A and 6A.
Unfortunately, we could not find the original data of the phospho-tyrosine and the DDR1 loading control blots outlined in Figure 5A of the paper. As requested, we therefore set up the method of DDR1 immunoprecipitation and performed the experiments as described for Figure 5A in the published paper. We used the same cell line as in the published paper and ascertained that the present batch of cells did not have an endogenous expression of Wnt-5a. We used the same Mastoparan and antibodies, although different batches had to be used. We pretreated the cells as described in the published paper and we treated the cells with vehicle and Mastoparan as previously described in the published paper. New and representative phospho-tyrosine and DDR1 loading control blots are included in the corrected Figure 5A. The blots are representative of six separate experiments. The results are essentially identical to those published.
We regret that we did not spot these mistakes earlier and we apologize for any inconvenience this may have caused.