Jingwu Zang says he is baffled by the whole affair. Until last month, he was head of a neurodegenerative-disease research unit in Shanghai, China, for London-based drug firm Glaxo­SmithKline (GSK). On 22 May, as he tells it, his boss told him that there would be an investigation. The next day, Chinese lawyers showed up at the company to interview him. On 31 May, he was told to hand in his computer and company credit card, and was escorted to his car. “Within a few minutes, I was outside the facility I built,” he says.

On 9 June, hereceived a letter informing him of his official termination of employment.

The investigation has focused on a paper published in Nature Medicine that Zang co-authored on multiple sclerosis (MS), his speciality (X. Liu et al. Nature Med. 16, 191–197; 2010). GSK is asking for the paper to be retracted; Zang stands by the results. The Chinese blogosphere is abuzz over the dispute, wondering what it signals for a centre seen as a bellwether for China’s budding drugs industry.

Zang set up the global research and development centre in Shanghai in 2007. The centre was considered bold: of the many inter­national pharmaceutical giants that had opened research operations in China in the previous five years, only GSK had given its branch wide autonomy, with control over global operations for an entire development sector, that of neuro­degenerative diseases. “In Shanghai, we can make decisions that drive global studies,” says Zang.

Now with some 400 scientific staff, the centre has several candidate neuro­degenerative drugs in phase I and II clinical trials, Zang says, and he was eager to get one through phase III, to “demonstrate that we can do great science and move a clinical compound forward”.

Four years ago, Zang’s group started work on a protein called the interleukin-7 receptor (IL-7R). “It was a really exciting story,” he says. IL-7R sits on the surface of certain immune cells, and a genetic variant of it had been linked to MS. Nobody knew what the underlying mechanism was, but Zang had a hypothesis — that the IL-7 pathway played a part in the pathogenic expansion of T-helper 17 (TH17) cells, immune cells that, when present in excess, are thought to contribute to MS.

In 2010, the group published results in Nature Medicine concluding that this was indeed the case. But last month, the paper came under scrutiny from within GSK after the company and Nature Medicine were notified of a problem with some of the data. A GSK investigation has since concluded that human blood samples used to create a figure in the article — described in the caption as being taken from patients with MS — actually came from healthy subjects.

On 10 June, GSK posted a statement saying: “Regretfully, our investigation has established that certain data in the paper were indeed misrepresented. We’ve shared our conclusion that the paper should be retracted and are in the process of asking all of the authors to sign a statement to that effect.”

Regretfully, our investigation has established that certain data in the paper were indeed misrepresented.

Zang and Xuebin Liu, the paper’s first author, both say that this was an unintentional mistake that does not change the paper’s overall conclusion. Liu, whose group ran the experiment and compiled the data, says that the team had hoped to use data from cells of patients with MS and had drafted a manuscript with that wording. But although preliminary data from patients did reveal Zang’s proposed link between the IL-7R pathway and TH17cells, staining in those images was inadequate — so the team turned to healthy subject data instead, Liu says. In a hurry to beat competition, they forgot to change the caption. Liu says that cells from either group can be used to show the effect.

Liu also addressed another problem, noted later on a pharmaceutical blog, after news of the investigation came out: two images, with captions describing different experimental conditions, are identical. Liu says that the mistaken duplication occurred during editing and layout of the article, and has asked Nature Medicine to check. The journal’s chief editor Juan Carlos López says that he cannot comment yet.

The main thrust of the paper — that IL-7R is related to MS, and that blocking its function can ameliorate MS-like disease in a mouse model — largely agrees with results from other groups. But scientists have failed to replicate the specific mechanism proposed by Zang’s team.

One of those studies, led by researchers at Stanford University in California and at Rinat, a subsidiary of the drugs giant Pfizer based in South San Francisco, California, found that the effects of blocking IL-7R largely agreed with the results from Zang’s group. But they were unable to reproduce the results of Zang and Liu’s experiments that supported a connection between IL-7R and TH17 cells as the mechanism. The discrepancy was “not likely due to differences in the experimental protocols, because we diligently followed their methods”, the researchers wrote in their 2011 report in Science Translational Medicine (L.-F. Lee et al. Sci. Transl. Med. 3, 93ra68; 2011). Stanford’s Lawrence Steinman, a corresponding author on the California study, declined to comment.

But Liu says that the California experiment differed from the China experiment in an important way: the GSK team used mature TH17 cells whereas the California group used undifferentiated ones. “It’s a different protocol, a different stage,” he says.

Liu says that as first author he takes full responsibility for the mistakes. On 9 June, he announced his resignation on a Chinese bioscience website. Both Liu and Zang say that they stand by the paper’s results and will not sign a letter to Nature Medicine requesting that the paper be retracted. López says that a retraction is still possible, even if not all of the authors agree to it, “if confidence in the paper is lost”. In such cases, “the paper is still retracted, explaining who agrees and who doesn’t agree to the retraction”, he says.

Asked whether a retraction is warranted if the mistakes do not affect the paper’s findings, David Daley, GSK’s director of global external communications, acknowledges in an e-mail that in the period since the research was carried out, “an independent body of evidence has accumulated that the receptor [for] interleukin-7 (IL-7) is a valid target for a variety of autoimmune disorders”. But he adds that “because certain data in the publication were misrepresented, we believe retracting the paper is the only appropriate action to take”.

Zang, who was involved in the experimental design and in drafting the manuscript, but not in the hands-on experiments or data compilation, believes that he was fired not over data misrepresentation but for allegedly “influencing the investigation”. The letter terminating his employment states that he “wilfully and purposefully undermined and misled this investigation and provided untruthful information” — charges he vehemently denies.

Daley declined to detail GSK’s reasons for terminating Zang’s employment, but provided a GSK statement: “We are confident in the thorough investigation we conducted and the actions we have taken as a result of our findings. We will not tolerate activity and behaviour that falls short of the high standards expected from our employees.”

Zang says the whole episode is bizarre. “I still can’t understand it.”

figure a


GlaxoSmithKline’s Shanghai unit has proved to be successful.