This study was approved by the Ethics committee at Karolinska Institutet, Etikkommitté syd, #689/03.
Cell lines and primary MCL
The MCL cell lines Jvm-2 [12, 13], Rec-1 [14, 15], Granta 519  and JeKo-1 , the precursor leukemia cell line Nalm-6  representing the pre-GC stage, the Burkitt's lymphoma cell lines Radji  and Namalwa  representing GC and the plasma cell line SKMM-2 [21, 22] representing post-GC were used in the study. All cell lines were tested for mycoplasma and were found to be free from infection. The tests were performed with MycoAlert Mycoplasma Detection Kit (LT07-218, Cambrex, USA) on a Thermo Luminoskan Ascent (Thermo Fischer Scientific, Inc, USA). The cell lines were cultured in RPMI 1640 GlutaMaxtm-1 (72400-021, Invitrogen, USA), supplemented with 10% FBS (16140-071, Invitrogen, USA) and 50 μg/mL gentamicin (15750-045, Invitrogen, USA) at 37°C using a 5% CO2 atmosphere, except JeKo-1, which was cultured using a 20% FBS supplement. The cells were harvested in the exponential phase, snap frozen in liquid nitrogen and stored as pellets at -70°C. In addition, two primary MCL tumors, with confirmed t(11;14)(q13;q32) cyclin D1 translocation and immunophenotype of CD5+/CD23-, indicative of MCL , were included in the study. The primary MCL tumours contained >80% tumor cells, as verified using the CD5 and CD23 markers in a FACScan (BD, USA) (not shown). Biological triplicates were used for the primary cells except MCL tumours, that were analyzed in duplicate due to lack of availability of samples.
Isolation of B cell populations from healthy individuals
The normal B cell populations were extracted from fresh tonsils obtained from patients undergoing tonsillectomies at the surgical department, Karolinska University Hospital, Huddinge. The tonsils were dissected and then homogenized using a Medimachine (Dako Denmark A/S, Glostrup, Denmark). The resulting cell suspension was purified from erythrocytes by treatment with a lysis buffer containing ammonium chloride for 8 minutes and subsequent filtration to remove cellular debris. The B cell population was separated from T cells by negative selection. The pan T cell CD3+ surface receptor was conjugated with magnetic anti-CD3+ microbeads (130-050-101, Miltenyi Biotec, Germany), which were sorted by an AutoMACS sorter (Miltenyi Biotec, Germany) using column 130-021-101 (Miltenyi Biotec, Germany). The purity of the T and B cell fractions was validated by using the pan B cell surface receptor CD19+
 and the T cell CD3+ receptor. The B cell fractions were >90% CD19+ as determined by incubating with fluorescent CD3+/FITC + CD19+/RPE antibodies (FR866, DAKO, Denmark) and analyzing with a FACScan (BD, USA) (not shown). The total B cell extracts were snap frozen in liquid nitrogen and stored as pellets at -70°C.
Isolation of B cell subpopulations representing pre-GC, GC and post-GC
In order to separate pre-GC, GC and post-GC populations from the tonsils, the total B cell populations from the AutoMACS separation were incubated with magnetic CD27+ microbeads (130-051-601, Miltenyi Biotec, Germany), generating a post-GC CD27+ population and a CD27- population consisting of both pre-GC and GC B cells. The CD27- population was then further sorted using an anti-IgD+-FITC antibody and Anti-FITC microbeads (130-048-701, Miltenyi Biotec, Germany), to isolate the pre-GC and GC fractions.
Protein preparation and concentration determination
Cell pellets from cell lines and isolated primary B cell populations were subjected to a lysis buffer (150 mM NaCl, 50 mM Tris-HCL pH 7,5, 0,1% Triton X-100, 1% CHAPS) and repeatedly immersed in liquid nitrogen to lyse the cells and release the proteins. Protein concentrations were determined using a BCA protein assay kit (23225, PIERCE, USA) according to the manufacturer's instructions.
Analysis of proteins using SELDI-TOF
The samples were dissolved to a final concentration of 0.15 ug/ul in 50 mM ammonium acetate pH 4.5 and 0.1% Triton X-100. 100 μl of the prepared protein solutions were analyzed using a cation exchange (CM 10, pH 4,5) protein-chip (Biorad Inc., USA) . Standard protocols recommended by the manufacturer were used in the experiments. Mass spectra were collected using SELDI-TOF Protein Biology System IIC using three different laser intensity and time lag focusing settings creating three data sets; 3-6 kDa, 6-10 kDa and 10-40 kDa (optimization of M/Z regions and of time lag focusing). The design of the experimental protocol was based on pilot experiments, to optimize peak number and minimize detector saturation from intense peaks at 5 kDa and 9 kDa. The data analysis was performed using CE software package Ciphergen Biosystems, CA, USA. All spectra were baseline corrected and peak detection was performed using signal/noise value of 3 and a valley depth of value of 2 (first pass).
The significant differentially expressed protein peaks were identified using nonparametric t-test, using in pair-wise comparison Mann-Whitney and in group-wise comparison Kruskal Wallis test. The peaks with p-value < 0,05 were considered as significant differentially expressed proteins peaks.
Clustering of protein profiles
Unsupervised hierarchical clustering of the primary populations and cell lines was performed using all peaks detected within each mass range. To distinguish between pre-GC, GC and post-GC populations, all peaks detected within the three populations which had p-value < 0,05 were considered significant and were used to cluster these primary populations within each data set. The corresponding peaks in the primary MCL samples and MCL cell lines were then included to generate a hierarchical clustering of MCL samples within the pre-GC, GC and post-GC populations. The peaks were analyzed by heat map and principal component analysis (PCA). Ciphergen Express software, Ciphergen, CA, USA, was used for both hierarchical clustering and PCA.