Jones and colleagues described 2 Hodgkin lymphoma (HL) cell lines that contain small subpopulations of aldehyde dehydrogenase (ALDH)+ CD20+CD30JIgλ+ B cells, which have clonogenic potential and give rise to the typical CD30+ Hodgkin and Reed-Sternberg (HRS) cells. 1 Importantly, the authors also reported that in the peripheral blood (PB) of HL patients, ALDHhigh cells are detectable, that are often clonally related to the HRS cells. These cells may represent HL-initiating cells, perhaps cancer stem cells. First indications for the existence of HRS cell clone members among small CD30J, morphologically normal cells were reported 10 years ago. In several HL the same numerical chromosomal abnormalities were present in HRS cells and in some small (CD30J) cells. 2, 3 However, such abnormalities are not the best clonal markers in HL. 4 In another study of HL with Epstein-Barr virus (EBV)+ HRS cells, CD30J small EBV+ lymphocytes in the HL lymph nodes were not (with perhaps rare exceptions) members of the HRS clone. 5 As HRS cells show a clonal EBV infection pattern, putative CD30J clone members should also be EBV-infected. Thus, members of the HRS cell clone are usually absent among CD30J B cells, at least in EBV+ cases. Vockerodt and colleagues analyzed multiple PB and bone marrow samples from 2 HL patients, using a highly sensitive HRS cell clone–specific PCR, but HRS cell–specific amplificates were undetectable. 6 Hence, clonotypic cells were very infrequent or absent in the PB of these patients.

Jones and coworkers present 3 pieces of evidence that clonotypic Bcells are present in the PB of HL patients among ALDHhighCD27+CD19+ B cells. 1 However, none of these is convincing. First, in the heavy chain fragment length analysis, the isolated HRS cell population surprisingly gave a polyclonal pattern with many fragments of different lengths, so that also seeing 2 bands with the same length as those obtained from the PB B cells does not prove clonal identity. Second, the fragment length analysis of Vκ light chain rearrangements is unsuitable to demonstrate clonal identity, because the VJ joints of light chain rearrangements show little length variation; 66% of polyclonal VκJκ joints have an identical CDRIII length of 27 bp. 7 Third, sequences of V (D) J