A case of alleged misconduct at the University of Washington in Seattle may finally be over. The Office of Research Integrity released its findings following an investigation into the work of Andrew Aprikyan, a former hematology researcher at the university.
The Aprikyan case has dragged on for a decade. In 2010, the university fired the scientist after a court denied his appeals based on allegations that they had denied him due process. As the Seattle Times reported at the time:
The case of Andrew Aprikyan, who is also the UW’s table tennis coach, shows just how difficult it can be for a university to determine whether a researcher made honest mistakes or fabricated results. At times, the case has pitted the UW’s administration against some faculty, who’ve supported Aprikyan.
It also raises questions about how the UW could let an academic-misconduct investigation drag on for so long, all the while allowing a medical researcher it suspects of wrongdoing to draw a salary, collect grants and travel the globe presenting his findings.
Aprikyan, according to the ORI — “Based on the report of an investigation conducted by the University of Washington (UW), the UW School of Medicine Dean’s Decision, the Decision of the Hearing Panel at UW, and additional analysis conducted by ORI” —
engaged in research misconduct in research supported by National Cancer Institute (NCI), National Institutes of Health (NIH), grant CA89135 and National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK), NIH, grant DK18951…
One paper — “Neutrophil Elastase Mutations in Severe Congenital Neutropenia Patients of the Original Kostmann Family,” which appeared online in Blood on January 16, 2003 — has been retracted and one, in Experimental Hematology , has been corrected. The ORI found problems in a third, also in Blood, but that paper does not have any corrections or retractions.
ORI’s description of Aprikyan’s wrongdoing — which he denies, according to the summary — is lengthy. ORI found that he:
falsely reported sequencing data in the NEM manuscript to strengthen the hypothesis that NE mutations contributed to the phenotype observed in severe congenital neutropenia (SCN) patients. Specifically:
a. Respondent falsely reported in Figures 2A and 3 that patient 3 had the R191Q neutrophil elastase (NE) mutation, when the majority of the sequencing experiments showed that the mutation was not present.
b. Respondent fabricated text (p. 12) reporting that sequencing of RT-PCR products confirmed the expression of the NE mutants in the SCN patients and that no mutations were present in the granulocyte colony stimulating factor receptor (G-CSFR) gene and the Wiskott-Aldrich Syndrome (WAS) gene in SCN patients, when based on the lack of original records the experiments were not performed. The false claim for G-CSFR sequencing was also reported in CA89135-03.
2. falsely reported a two-fold increase in apoptosis of human promyelocytic (HL-60) cells transfected with NE mutants compared to wild type NE in Figure 4A, NEM, Figure 6A, CMA, Figure 8, HL73063-01, and Figure 7, HL79615-01. Respondent used arbitrary flow cytometry data files to generate histograms with the desired result. The false results supported the hypothesis that the NE mutations were sufficient for impaired survival of human myeloid cells.
3. falsified NE and ß-actin Western blots in Figure 4B Blood, pre-published online January 16, 2003, Figure 5B of the manuscript initially submitted to Blood April 2002, and Figure 6B Experimental Hematology 31:372-381, 2003, by falsely labeling lanes to support the hypothesis that accelerated apoptosis in mutant NE transfect HL-60 cells was due to the mutation and not the level of protein present. Specifically:
a. Respondent used portions of a single NE Wester blot to represent: Figure 4B as HL-60 cells transfected with L92H, R191Q, and wtNE, when the cells were transfected with R191Q, P110L, and D145-152; Figure 5B as HL-60 transfected with wtNE, mutNE, and EGFP when they were cells transfected with NE mutants, P110L, D145-152, and 194
b. Respondent used portions of a single ß-actin Western blot to represent: Figure 4B as HL-60 cells transfected with L92H, R191Q, and wtNE, when they were cells transfected with I31T, P110L, and G185R mutants; Figure 5B as HL-60 cells transfected with wtNE, mutNE, and EGFP, when they were cells transfected with P110L, I31T, and INE; Figure 6B as HL-60 cells transfected with G185R, mock, D145-152, and P110L NE mutants, when they were cells transfected with I31T, P110L, G185R, and 32. The false ß-actin Western blot in Figure 6B was also included in HL73063-01, Figure 8 (where the I31Tlane was labeled correctly), and HL79615-01, Figure 7.
4. falsified the reported methodology for flow cytometry experiments in Figure 4A, NEM, Figure 1 and 2, and Tables 2 and 3, CMA, and Figures 4, 5, and 6, ISB, to validate the key hypothesis showing accelerated apoptosis in SCN and CN patients. The methodology claimed that flow cytometry experiments were gated for GFP+ populations, or that cell purity was greater than 96%, when based on the available original records, the experiments were not performed as stated.
5.falsified Figure 2, CMA, Figure 2, HL73063-01, Figure 3, HL79615-01, and Figure 5, CA89135-01A1, demonstrating that the overnight cultures of CD34+ and CD33+ bone marrow cells from SCN/AML patients showed normal cell survival, and only the CD15+ overnight cultures showed accelerated apoptosis, when the actual record available contradicted this result. Respondent used flow cytometry data files to generate histograms with the desired result to support the hypothesis that the progression from SCN to leukemia (AML) involves acquired G-CSFR mutations that override the pro-apoptotic effect of the NE mutations in primitive progenitor cells.
Aprikyan agreed to have any federally-funded work supervised for two years, and not to serve on any Public Health Service committees — eg NIH peer review committees — for the same amount of time, beginning March 12, 2013.