At team of researchers at MD Anderson Cancer Center has retracted a paper after realizing that the cell lines they were using weren’t what they thought they were.
Here’s the detailed notice:
The authors wish to retract the article titled “Critical Role of Notch Signaling in Osteosarcoma Invasion and Metastasis,” which was published in the May 15, 2008, issue of Clinical Cancer Research (1).
Many years after this article was published, we obtained new information showing that two of the cell lines used in this publication, OS 187 and COL, are not osteosarcoma cell lines. Both lines were first reported by us (2).
Each of these cell lines has now been tested by DNA fingerprint analysis. OS 187 tested identically to the common NCI60 colon cancer line HCT 15, even in a vial from passage 2 in our laboratory that had been frozen in 2002. Injection of OS 187 into the cecum of NOD/SCID/IL-2Rγ−/− mice yielded a colonic primary with metastasis to the liver (data not shown). Therefore, all of our published results with OS 187 likely reflect the biology of colon cancer. Given the subsequent report by another group that Notch pathway signaling promotes invasion and metastasis of colon cancer (3), the data we reported for OS 187 are not surprising.
The pattern observed for tumor growth after injection of COL cells into the tibia of NOD/SCID/IL-2Rγ−/− mice (no bone primary, rapid growth of liver tumors, with additional masses arising in skin and retro-orbital spaces) suggested that this cell line likely represented neuroblastoma. We have recently reported the subsequent testing confirming that COL is a neuroblastoma (4).
In this article, OS 187 and COL were included in Figs. 1 and 2. OS 187 was among the cell lines shown in Fig. 3 and is the only cell line shown in Figs. 4 and 5. When these two cell lines are removed from the data presented, two findings reported should be interpreted with caution. First, the only real osteosarcoma cell lines in which Notch was directly manipulated genetically were SAOS2 (with low levels of Notch) and LM7 (a subline of SAOS2 with high levels of Notch). Although transduction of SAOS2 with constitutively active intracellular Notch1 (ICN1) or the Notch target gene Hes1 increased in vitro invasiveness, LM7 was not able to grow after transduction with these constructs, as indicated in the original publication. Thus, we have not proved that osteosarcoma invasiveness can be increased by further upregulating Notch activity in Notch-positive osteosarcoma. Second, because only OS 187 was used in vivo in this report, we can no longer say that our publication proves Notch’s role in metastasis in vivo for osteosarcoma. However, the subsequent confirmation of this function of Notch by other groups (5) renders it likely that this conclusion remains valid for osteosarcoma. Nonetheless, because the data provided from bona fide osteosarcoma cell lines are no longer adequate to support the assertions and conclusions of the article, the authors respectfully request its retraction.
Pingyu Zhang
Yanwen Yang
Patrick A. Zweidler-McKay
Dennis P.M. Hughes
This sort of cell line mixup has been the reason for several retractions by unrelated groups, including some because of contamination by the HeLa cell line.
The paper has been cited 76 times, according to Thomson Scientific’s Web of Knowledge. Kudos to Hughes and his colleagues for coming forward and writing such a detailed notice.
This extremely clear and detailed retraction is impressive and honorable. What a great example of how things should be done.
Yes, indeed. One common, though not universal feature of retractions such as this on that fall into the “Doing the right thing” category is that the authors spend time and resource trouble shooting, to figure out what the problem is.