Exportin 1‐mediated nuclear/cytoplasmic trafficking controls drug sensitivity of classical Hodgkin's lymphoma

Exportin 1 (XPO1) is the main nuclear export receptor that controls the subcellular trafficking and the functions of major regulatory proteins. XPO1 is overexpressed in various cancers and small inhibitors of nuclear export (SINEs) have been developed to inhibit XPO1. In primary mediastinal B‐cell lymphoma (PMBL) and classical Hodgkin's lymphoma (cHL), the XPO1 gene may be mutated on one nucleotide and encodes the mutant XPO1E571K. To understand the impact of mutation on protein function, we studied the response of PMBL and cHL cells to selinexor, a SINE, and ibrutinib, an inhibitor of Bruton tyrosine kinase. XPO1 mutation renders lymphoma cells more sensitive to selinexor due to a faster degradation of mutant XPO1 compared to the wild‐type. We further showed that a mistrafficking of p65 (RELA) and p52 (NFκB2) transcription factors between the nuclear and cytoplasmic compartments accounts for the response toward ibrutinib. XPO1 mutation may be envisaged as a biomarker of the response of PMBL and cHL cells and other B‐cell hemopathies to SINEs and drugs that target even indirectly the NFκB signaling pathway.

Table S4.Sequences of the primers used for XPO1 PCR amplification and Sanger sequencing Forward 5'-TGT GTT GGG CAA TAG GCT CC-3' Reverse 5'-TGA ACC TGA ACG AAA TGC CTG C-3' The primer sequences were designed with the primer 3 software (v4.0, //primer3.ut.ee/).The primer sequences directed against the corresponding genes were designed with the primer 3 software (v4.0, //primer3.ut.ee/).GAPDH and RPLP0 served as internal controls.24.8At the end of the experiment, tumors were measured and their volumes calculated with the formula: volume in mm 3 = 1/2(length × width 2 ) using digital calipers.As before, the volumes of tumor were highly heterogeneous.Due to this heterogeneity and the low number of animal in each series, only the combo vs vehicle p-value is statitically diffreent with p = 0.024 (see Fig. 3C).

Table S14. Number of IPO1/XPO1 or IPO1/NFκB dimers in cHL cells
From the PLA assays, the number of red dots representing one dimer was counted in the indicated number of cells (n) with the ImageJ software.The number of the minimum and maximum dots in one cell was recorded, the means as well as s.d. were calculated with the same software.

Fig. S1. Schematic representation of the CRISPR-Cas9 knockout strategy set up for MedB1 cells
We used an sgRNA (Table S3, green arrow) that targets the E571 codon (in yellow and bold) to delete the wt XPO1 allele.To ensure a high editing efficiency, we choose the PAM site (in grey) close to the targeted codon.MedB1 cells were transfected by nucleofection.Two days later, gDNA was purified from edited MedB1 cells, PCR-amplified with the primers described in the Table S4.PCR fragments were purified and sequenced by the Sanger method.Edited cells (Δwt) were maintained in culture and compared to parental cells (p).The expression of XPO1 was analyzed by indirect IF (Fig. 2E).The sensitivity of both cell types toward selinexor and ibrutinib was assessed with an MTS assay (Fig. 2F).Membranes were cut into strips that were incubated with specific Abs recognizing either poly (ADP-ribose) polymerase 1 (PARP1) as nuclear protein or enolase 1 (ENO1) as cytosolic protein (Table S1).S1).Three independent blots were run and each blot was analyzed by densitometry for the quantification of XPO1 protein level relative to the β-actin level used as an internal control.The level of XPO1 in selinexor-treated cells was then normalized to the one in vehicle-treated cells defined as 1.The values were used to draw the histogram presented Fig. 1E and for the statistical analyses.

Fig. S4. Western blots for controlling the efficacy of ibrutinib and selinexor treatments
L428 and UHO1 cells were treated with selinexor or ibrutinib or the combination of the two drugs for 6 or 24 h.Whole-cell proteins were purified.Proteins (30 μg) separated by SDS-PAGE, and transferred onto nitrocellulose membranes stained with Ponceau red.According to the scheme, membranes were then cut into strips that were incubated with the indicated Abs (Table S1).Selinexor treatment leads to the degradation of XPO1 in both cell lines 24 h post-treatment, whereas the ibrutinib treatment leads to the disappearance of the phosphorylated form of BTK at the same time.An anti-β-actin Ab served as a control of charge and transfer.Molecular weight markers (MW) were run in the same gels.

Fig. S5. U2940 cells, KS/KAS and EAS/ES derivatives display the same proliferation curve
Cells were seeded at the density of 3 x 10 5 cells/ml in 24-well plates in complete medium and counted by trypan blue exclusion every day for one week (n = 3).During this period, cells were not diluted.The means of total number of living cells are reported on the graph together with s.d.The experiment was done twice, a representative one is presented.

Fig. S6. Characterization of the in vivo/in ovo model of cHL cell lines engraftment
A, Fertilized eggs (D1, EARL Les Bruyères, Dangers, France) were incubated for nine days at 38°C in an incubator with 55% relative humidity.At that time, the CAM was dropped by drilling a small hole through the eggshell into the air pocket and a small window was cut in the eggshell above the CAM (D9).Cultured tumor cells (10 6 ) mixed in Matrigel (v/v, 50 μl final volume) were directly added on the CAM beneath.Tumors were visible as soon as two days post-engraftment (D11) and grew all along the duration of the experiment (D16).The vascularized tumor is arrowed in white.B, At the end of the experiment (D16), the upper CAM was carefully removed from each egg and the tumor cut out.Tumors were weighted and tumor cells were dissociated with Accutase (Merck) then analyzed for CD15 expression by flow cytometry.Ninetyseven % of cells were CD15-positive in the culture of L428 cells and 85% in the tumor after cell dissociation.C, Dissociated tumor cells were cytospinned onto glass slides and stained with May-Grünwald-Giemsa for morphological examination (X 1 000, magnification).

Fig. S8. Effects of drugs on L428 tumor weights
Ten millions of L428 cells mixed in matrigel were engrafted s.c.onto the flank of SCID mice.Eight weeks later, tumor-bearing mice were separated into fours groups receiving : vehicle (n = 4), ibrutinib (I), selinexor (S) or the combination of drugs (C) (n = 5, in each arm).Two mice died in the ibrutinib group.Thirty-five days after the beginning of the treatments, mice were sacrificed and the tumors recovered.They were weightedand the volume of each tumor in each treatment arm is reported in the graph together with the statistical analysis.* p < 0.05 with the t-test.A similar graph was drawn with the tumor volumes (Fig. 3C).

Fig. S10. DNA binding activity of NFκB proteins in cHL
The DNA binding activity of cHL nuclear protein extracts was quantified with an ELISA-based assay (TransAM NFκB Activation Assay Kit, Active Motif, Carlsbad, CA).Nuclear proteins were prepared from cultured cells using the Nuclear Extract Kit (Active Motif) then quantified.The assay was performed with 2 μg of nuclear proteins, each condition was set up in triplicate according to the manufacturer's protocol.
The assay was run twice and analyzed using a multilabel plate reader (Victor X4, Perkin Elmer, Waltham, MA).The binding activity of each NFκB family member (p50, p52, p65, cREL, RELB) is reported on the histogram for L428, L1236 and UHO1 cHL cell lines as well as Raji cells, the internal positive control (means ± s.d.).

Fig. S11. Indirect immunofluorescence analysis of NFκB proteins in cHL cells
NFκB proteins expression was analyzed by IF in the four cHL cell lines.We used primary Abs (Table S1) and a goat Alexa Fluor 633-conjugated anti-mouse IgG or anti-rabbit IgG as secondary Abs.Slides were counterstained with DAPI and analyzed by confocal microscopy (x 180, magnification).Representative enlarged images (x 4) are shown.FIs of Alexa-633 fluorophore (in red) and DAPI (in blue) were estimated with the ImageJ software and data were exported to generate the curves of FI as a function of the distance along the white axis crossing the cell.

Fig. S12. Western blot analysis of p-p65 expression in cHL cell lines
Nuclear and cytoplasmic extracts were sequentially purified from cultured cHL cells treated with vehicle or MG132 (1 μM for 24 h).Cytosolic proteins (30 μg) were separated by SDS-PAGE, transferred onto nitrocellulose membranes.Membranes were cut and strips were then incubated with the indicated Abs (Table S2).Three independent blots were run and each blot was analyzed by densitometry for the quantification of p-p65 protein level relative to the β-actin level used as an internal control.The level of p-p65 in MG132-treated cells was then normalized to the one in vehicle-treated cells defined as 1.The values were used to draw the histogram presented in Fig. 6B and for the statistical analyses.

Fig. S13 -Original blots for Fig. 2C
Whole-cell proteins were purified, separated on SDS-PAGE, tranferred onto nitrocellulose membranes that were stained with Ponceau red.As presented membranes were cut into strips and incubated with the mentionned Abs (Table S1).Molecular weight markers (MW) were run in the same gels.
Fig. S2.Sequential extraction of cytoplasmic and nuclear proteins Cultured cHL cells were harvested.Cytosolic and nuclear extracts were prepared with the NE-PER Nuclear and Cytoplasmic Extraction Reagent (Thermo Scientific) according to the manufacture instructions.Proteins (30 μg) were separated on SDS-PAGE and transferred onto nitrocellulose membrane.Membranes were cut into strips that were incubated with specific Abs recognizing either poly (ADP-ribose) polymerase 1 (PARP1) as nuclear protein or enolase 1 (ENO1) as cytosolic protein (TableS1).

Fig. S3 .
Fig. S3.Degradation of the XPO1 protein in cHL expressing the wt or the mutant XPO1 protein cHL cell lines were either treated with vehicle-(Vh) or selinexor-(S) (10 or 100 nM) for 8 or 24 h.Whole-cell proteins were purified and separated on SDS-PAGE, transferred onto nitrocellulose membranes, then incubated with the indicated Abs (TableS1).Three independent blots were run and each blot was analyzed by densitometry for the quantification of XPO1 protein level relative to the β-actin level used as an internal control.The level of XPO1 in selinexor-treated cells was then normalized to the one in vehicle-treated cells defined as 1.The values were used to draw the histogram presented Fig.1Eand for the statistical analyses.

Table S2 .
B-cell lymphoma cell lines characteristics

Table S7 . Effects of drugs treatment on L428 tumor weight (mg) Vehicle Ibrutinib Selinexor Combo
Two mice died unexpectedly during the course of the treatment in the ibrutinib series.To illustrate the high tumor viariability among the mice, the minimum, maximum, and range of tumor weights are indicated in the table as well as the means ± s.d.The p-values calculated with the t-test, were not statistically different except p = 0.0112 for combo vs vehicle and p = 0.0471 for combo vs selinexor (see Fig.S8).

Table S9 . Fluorescence intensity of nuclear NFκB proteins in cHL
Cultured cHL were analyzed by indirect IF for NFκB proteins (NFκB1/p50, NFκB2/p52, p65 (RELA), RELB and cREL) localization and expression.The minimum, maximum and mean of FI for nuclear proteins are indicated in the table as well as the number of individual cells recorded (n) for each experimental condition.The corresponding boxplots are presented in Fig.4D.

Table S10 . Fn/c calculated from the four cHL cell lines following selinexor treatment
Cultured cells were treated with vehicle (V) or selinexor (S), then analyzed by indirect IF for NFκB proteins localization.The means of calculated Fn/c for each antibody are indicated in the table as well as the number of individual cells recorded (n) for each experimental condition.We calculated the ratio Fn/c with the ImageJ software.An increased Fn/c indicated the nuclear accumulation of the protein.The cooresponding boxplots are reported Fig.5A.To compare selinexor-vs vehicle-treated cells, the p-value (paired t-test) was calculated with the PRISM software.

Table S11 . Fluorescence intensity of nuclear nucleophosmin and survivin in SUPHD1 and UHO1 cells treated with selinexor
Cultured SUPHD1 and UHO1 cells were treated with vehicle or selinexor (100 nM for 6 h), then analyzed by indirect IF for nucleophosmin (NPM) and survivin (BIRC5) localization.The minimum, maximum and mean of fluorescence intensity (FI) for nuclear NPM and BIRC5 are indicated in the table as well as the number of individual cells recorded (n) for each experimental condition.The corresponding boxplots are presented in Fig.5B.

Table S12 . Fn/c calculated from L428, L1236 and UHO1 cell lines following importazole treatment Cell line
Cultured cells were treated with vehicle (V) or importazole (I), then analyzed by indirect IF for NFκB proteins localization.The mean of FI for each antibody is indicated in the table as well as the number of individual cells recorded (n) for each cell line.We calculated the ratio Fn/c with the ImageJ software as reported previously.They are presented in the Fig.5C.The p-value (paired t-test) was calculated with the PRISM software for the comparison of importazole-vs vehicle-treated cells.

Table S13 . Fluorescence intensity of cytoplasmic p65 protein in cHL
Cultured cHL were analyzed by indirect IF for nuclear p65 expression.The minimum, maximum and mean of fluorescence intensity (FI) of cytoplasmic proteins is indicated in the table as well as the number of individual cells recorded (n) for each cell line.The corresponding boxplots are presented in Fig.6A.