Deletion of murine Rhoh leads to de-repression of Bcl-6 via decreased KAISO levels and accelerates a malignancy phenotype in a murine model of lymphoma

ABSTRACT

RHOH/TFF, a member of the RAS GTPase super family, has important functions in lymphopoiesis and proximal T cell receptor signalling and has been implicated in a variety of leukaemias and lymphomas. RHOH was initially identified as a translocation partner with BCL-6 in non-Hodgkin lymphoma (NHL), and aberrant somatic hypermutation (SHM) in the 5ʹ untranslated region of the RHOH gene has also been detected in Diffuse Large B-Cell Lymphoma (DLBCL). Recent data suggest a correlation between RhoH expression and disease progression in Acute Myeloid Leukaemia (AML). However, the effects of RHOH mutations and translocations on RhoH expression and malignant transformation remain unknown. We found that aged Rhoh^−/− (KO) mice had shortened lifespans and developed B cell derived splenomegaly with an increased Bcl-6 expression profile in splenocytes. We utilized a murine model of Bcl-6 driven DLBCL to further explore the role of RhoH in malignant behaviour by crossing Rhoh^KO mice with Iµ-HABcl-6 transgenic (Bcl-6^Tg) mice. The loss of Rhoh in Bcl-6^Tg mice led to a more rapid disease progression. Mechanistically, we demonstrated that deletion of Rhoh in these murine lymphoma cells was associated with decreased levels of the RhoH binding partner KAISO, a dual-specific Zinc finger transcription factor, de-repression of KAISO target Bcl-6, and downregulation of the BCL-6 target Blimp-1. Re-expression of RhoH in Rhoh^KOBcl-6^Tg lymphoma cell lines reversed these changes in expression profile and reduced proliferation of lymphoma cells in vitro. These findings suggest a previously unidentified regulatory role of RhoH in the proliferation of tumour cells via altered BCL-6 expression. (250)

KEYWORDS:

• RhoH
• KAISO
• BCL-6
• Blimp-1
• lymphoma
• malignant transformation
• B cells

Introduction

Rho GTPases are guanine nucleotide-binding proteins that cycle between an active GTP-bound state and an inactive GDP-bound state. As such, they have key roles in the modulation of a wide range of cellular processes including proliferation, apoptosis, cell migration, cell polarization, membrane trafficking, cytoskeleton rearrangements, and transcriptional regulation [1].

RhoH, a haematopoietic-specific and GTPase-deficient RhoGTPase, was first identified as a hypermutable gene and translocation partner in human Diffuse Large B-Cell Lymphoma (DLBCL) and was initially named Translocation Three Four (TTF) [2–4]. This atypical Rho GTPase lacks intrinsic GTPase activity and remains in a constitutively active, GTP-bound state [2,5]. In addition, it contains a functional immunoreceptor tyrosine-based activation motif (ITAM)-like motif [6]. Thus, RhoH cellular function is likely related to its level of expression, intracellular localization, and phosphorylation state [6–11]. RhoH antagonizes Rac1 in regulating F-actin assembly and chemotaxis in haematopoietic stem/progenitor cells (HSPCs) by reducing Rac1 membrane targeting and activation, and this regulation is mediated by C-terminal sequences, including a CKIF motif, which is a C-terminal RhoH-specific motif of four amino acids (Cysteine, Lysine, Isoleucine, and Phenylalanine) [12]. Rhoh-deficient (Rhoh^KO) haematopoietic cells exhibit abnormal random migration [7]. In T cells, RhoH is required for CD3-zeta phosphorylation and recruitment of the protein tyrosine kinases Zap70 and Lck to the cellular membrane and immunologic synapse [6]. Rhoh^KO mice exhibit a T cell deficiency due to impaired T cell receptor-mediated selection and maturation of thymocytes. However, these mice show only minor alterations in B cell development and in vitro function [9].

RHOH mutations have been reported in 35 of 100 DLBCL cases [13] and mutations of the 5ʹ-noncoding region and first intron of the RHOH gene have been found in a variety of B lymphoid malignancies, including DLBCL [4]. In particular, aberrant somatic hypermutation (SHM) in the 5ʹ untranslated region of the RHOH gene has been detected in DLBCL. However, the effect of RHOH mutations and translocations on RhoH expression and the biology of lymphoma are currently unknown. Recent findings of abnormal RhoH expression levels in haematological malignancies further suggest that it may play an important role in malignant transformation and progression. Under-expression of RhoH has been shown to correlate with disease progression and decreased survival in Acute Myeloid Leukaemia (AML) [12], while constitutively active Rac1 has been demonstrated to inhibit apoptosis in human lymphoma cells by stimulating Bad phosphorylation [14]. Therefore, it is possible that decreased expression of RhoH may lead to increased protection of leukaemia cells from apoptosis via Rac1 inhibition [12] and contribute to chemotherapy resistance.

In this study, we demonstrate that Rhoh^KO mice have decreased lifespans and develop features of lymphoma including enlarged spleens with expanded CD19⁺ B cell populations that exhibit increased levels of BCL-6 and Ki-67 expression when compared with wild type (WT) mice. To better define the effects of dysregulated RhoH on lymphoma development, we generated Rhoh-deficiency in a DLBCL mouse model by interbreeding Rhoh^KO mice with Iµ-HABcl-6, Bcl-6^Tg mice that constitutively express BCL-6 in B cells. We report that loss of RhoH is associated with a more aggressive disease phenotype in this mouse model and Rhoh-deficient Bcl-6^Tg lymphoma cells show a higher proliferative activity compared with WT cells. Mechanistically, we demonstrate binding of RhoH and KAISO, a dual function zinc finger transcriptional repressor of Bcl-6, and show decreased cellular content of KAISO in the absence of RhoH in lymphoma cells. This decreased protein level is associated with de-repression of endogenous Bcl-6 and downregulation of the BCL-6 target gene Blimp-1, deficiency of which is associated with an aggressive phenotype in DLBCL. Our results therefore contribute new understanding to the potential role of deregulated RhoH expression levels in the development of Bcl-6-driven lymphoma.

Materials and methods

Mice

The generation of the Rhoh^KO mouse model on a C57BL/6J background has been described previously [6]. The Bcl-6^Tg mouse model with a C57BL/6 J background was kindly provided by Dr. Laura Pasqualucci. The Rhoh^WTBcl-6^Tg (WTTg) animals were generated by breeding Rhoh^WT mice with Bcl-6^Tg mice, and the Rhoh^KOBcl-6^Tg (KOTg) mice were generated by breeding Rhoh^KO mice with Bcl-6^Tg mice.

Vector constructs

The mouse stem cell virus (MSCV)–based retroviral vector with yellow fluorescent protein (YFP) and 3xhaemagglutinin-(HA)-tagged wild-type mouse RhoH cDNA (HA–RhoH) [15] was used to transduce Rhoh^KOBcl-6^Tg (KOTg) B cell lymphoma and Rhoh^KOBcl-6^nTg (KOnTg) B cell malignant cell lines. A FLAG-tagged wild-type mouse Blimp-1 cDNA was subcloned into an enhanced green fluorescent protein (EGFP) expressing modified MSCV-based retroviral vector, MIGR1 (Addgene Plasmid # 27,490, Watertown, MA), at XhoI and EcoRI restriction sites. This vector (FLAG-BLIMP-1) was used for the transduction of the KOTg B cell lymphoma cell line (see Supplemental Figure S1).

Establishment of murine lymphoma cell lines, cell culture and transduction

To establish Rhoh^KOBcl-6^Tg (KOTg) lymphoma cell lines with both endogenous and exogenous BCL-6 expression and a Rhoh^KOBcl-6^nTg (KOnTg) malignant B cell line with lymphoma characteristics having only endogenous BCL-6 expression, CD19⁺ cells were isolated from enlarged spleens of animals diagnosed with B cell lymphoma or with enlarged spleens and expanded CD19⁺ cells, respectively, by pathological examination. The cells were then cultured in RPMI1640 (Cytiva Hyclone Laboratories, South Logan, UT) with 20% foetal bovine serum (Sigma, St. Louis, MO) and 100 mM of 2-mercaptoethanol (Sigma) in vitro. After 2–3 weeks in culture, these lymphoma and B cell-derived cell lines were cryopreserved and a subculture passaged serially. The surface expression of CD19 on these cell lines was confirmed by fluorescence activated cell sorter (FACS) analysis.

Histology preparations

Spleens from mice were formalin-fixed and paraffin-embedded. On serial histological sections, haematoxylin and Eosin (H&E) staining and immunohistochemistry staining were performed at the Rodent Histopathology Core Facility or Brigham and Women’s Hospital Pathology Core Facility using standard protocols.

Flow cytometry analysis

For flow cytometry analysis, the following fluorescence-labelled antibodies were used: Ki-67-PE (20 µL per reaction), which was purchased from BD Biosciences, San Jose, CA, and CD19-FITC (6D5, 1:100), which was obtained from BioLegend, San Diego, CA. Ki-67 staining was conducted according to the manufacturer’s recommendations. Cells were analysed on a LSRFortessa flow cytometer (BD Bioscience) using BD FACSDiva software.

Immunoprecipitation and immunoblotting

Primary CD19⁺ B cells of wild-type non-transgenic (WTnTg), wild-type transgenic (WTTg), and Rhoh KO transgenic (KOTg) spleens were isolated using magnetic beadsconjugated with CD19 antibody (90 µL per 10⁷ cells) and an autoMACS Pro separator (Miltenyi Biotec, Bergisch Gladbach, Germany) according to the manufacturer’s recommendations. The purity of the isolated CD19⁺ cells was more than 90% as determined by flow cytometry.

To obtain whole-cell lysates, RIPA buffer (10 mM TrisHCl pH 7.4, 130 mM NaCl, 1% Triton X-100, 0.1% sodium dodecyl sulphate (SDS), 0.5% sodium deoxycholate) containing protease (Complete Protease Inhibitor Cocktail, Roche Applied Science, Indianapolis, IN) was used to lyse the isolated cells. The cytoplasmic and nuclear proteins were extracted separately using NE-PER nuclear and cytoplasmic extraction reagents (Pierce Biotechnology, Rockford, IL, USA) according to the manufacturer’s protocol.

For immunoprecipitation, the cells were lysed in 1X cell lysis buffer (20 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1 mM Na₂EDTA, 1 mM EGTA, 1% Triton, 2.5 mM sodium pyrophosphate, 1 mM β-glycerophosphate, 1 mM Na₃VO_4, 1 µg/ml leupeptin) (Cell Signalling Technology, Danvers, MA). The cell lysates were subjected to immunoprecipitation with anti-HA monoclonal antibody (1:100) (Thermo Scientific, Waltham, MA) followed by Protein A Magnetic Bead isolation (Cell Signalling Technology).

Western blotting was done following standard procedures and transferred onto PVDF membranes. The following primary antibodies were used: anti-BCL-6 (Cell Signalling Technology, 1:1000), anti-BLIMP-1 (BioLegend, 1:100 or Cell Signalling Technology, 1:250), anti-KAISO (Abcam, Cambridge, MA, 1:1000), anti-HA (Thermo Scientific, Waltham, MA, 1:1000), anti-Lamin B (Abcam, 1:1000), anti-Tublin A (Abcam, 1:1000), and anti-β-actin (Thermo Scientific, Waltham, MA, 1:5000 or 1:20,000). The sheep anti-mouse IgG (1:3000) or the anti-rabbit IgG (1:3000) horseradish peroxidase linked antibodies from Cell Signalling Technology were used as secondary antibodies and ECL reagent (Super Signal West Pico, Thermo Scientific, Rockford, IL) was used for detection.

Indirect immunofluorescence

Cells were layered onto poly-L-lysine coated slides and subsequently fixed with 4% paraformaldehyde and permeabilized with PBS–0.2% Triton X-100 for 5 minutes at room temperature. Fixed cells were blocked with 2% bovine serum albumin (Sigma) in PBS at room temperature for 1 hour, then incubated with anti-KAISO antibody (Abcam, 1:500) overnight at 4°C, followed by a 50-minute incubation with Alexa Fluor® 647 (1:500) conjugated anti-mouse IgG antibody (Cell Signalling Technology). Coverslips were mounted with ProLong Glass Antifade Mountant with NucBlu Stain (Thermo Fisher Scientific, Waltham, MA), and images were captured using a laser confocal microscope (Leica TCS SP8 (STED One) Scanning Confocal-Microscope, Wetzlar, Germany) with 100X magnification.

Chromatin immunoprecipitation (ChIP-qPCR)

ChIP was performed as previously described [16]. Briefly, cells were fixed in 1% formalin (v/v) in PBS with gentle rotation for 10 minutes at room temperature. Fixation was stopped by the addition of 125 mM glycine. Fixed cells were washed twice in ice-cold PBS and resuspended in sodium dodecyl sulphate (SDS) lysis buffer (1% SDS, 10 mM EDTA, 50 mM Tris-HCl, pH 8.1). Chromatin was sheared by sonication to about 100–500bp fragments using Bioruptor (Diagenode, Denville, NJ) and diluted ten-fold with dilution buffer (0.01% SDS, 1.1% Triton-X100, 1.2 mM EDTA, 16.7 mM Tris-HCl, pH8.1,167 mM NaCl). Antibody against KAISO (Abcam, 8 µg per reaction) was used to precipitate DNA fragments associated with KAISO. Precipitates were washed sequentially with ice cold low salt wash (0.1% SDS, 1% Triton-X-100, 2 mM EDTA, 20 mM Tris-HCl, pH8.1, 150 mM NaCl), high salt wash (0.1% SDS, 1% Triton-X-100, 2 mM EDTA, 20 mM Tris-HCl, pH 8.1, 500 mM NaCl), LiCl wash (0.25 M LiCl, 1% IGEPAL CA-630, 1% deoxycholic acid, 1 mM EDTA, 10 mM Tris-HCl, pH 8.1), and TE wash (1 mM EDTA, 10 mM Tris-HCl, pH 8.1) and finally eluted in elution buffer (1% SDS, 0.1 M NaHCO3). All buffers except for the elution buffer were supplemented with protease inhibitors (Complete Mini Protease Inhibitor Cocktail tablets, Roche). Eluted DNA fragments were analysed by quantitative PCR using promoter specific primers as listed.

RNA extraction and real-time PCR

Trizol (Invitrogen, Carlsbad, CA) was used to extract RNA from viable cells. Polymerase chain reaction (PCR) master mixes were used (SYBR Green, Applied Biosystems, Grand Island, Ny), and real-time PCR was performed using the 7500 Real-Time PCR System (Applied Biosystems). Further assay details can be provided on request.

Statistics

Microsoft Excel and GraphPad Prism software were used for statistical analysis. Statistical significance between two groups was determined by unpaired 2-tailed Student’s t-test. The Kaplan–Meier method was used to plot survival curves for murine B cell lymphoma mice, and the log-rank test was used to evaluate statistical differences. A p-value less than 0.05 was considered significant. Image J software from NIH was used for Western blot quantification.

Study approval

All animal experiments described in this study were approved by and adhered to the guidelines of the Institutional Animal Care and Use Committee of Boston Children’s Hospital. (Protocol number: 17–01-3365)

Results

Rhoh^−/− (KO) mice have shortened lifespans and splenomegaly due to an expansion of proliferating B cells with increased BCL-6 expression

Although Rhoh deletion (KO) had only a modest effect on B cell development in mice [9], we observed these animals had shortened lifespans when compared to WT mice (562 days vs 692 days; KO vs WT, median survival, p < 0.01) (Figure 1(a)) that was associated with increased spleen weight in KO mice early in life (Figure 1(b)). Spleen histology showed prominent expansion of the white pulp areas with effacement of the structure due to a diffuse proliferation of medium-large lymphoid cells in KO mouse spleens. Immunohistochemistry demonstrated that the expanded cells were mostly CD19⁺ B cells with increased percentages of BCL-6 and Ki-67 positive cells in the absence of Rhoh when comparing WT and KO mouse spleens (Figure 1(c)). The abnormal spleen architecture and the significantly increased percentage of Ki-67⁺ cells in the KO mouse spleen compared to the WT control (Figure 1(d)) suggest that splenomegaly was likely due to increased proliferation of B lymphoma cells. The role of RhoH in the increased proliferation was confirmed using a B cell line derived from the enlarged spleen of this KO mouse. RhoH was re-expressed in these cells using retrovirus-mediated gene transfer. RhoH re-expression significantly decreased cell proliferation of these KO mouse cells (Figure 1(e)). These data imply that RhoH deficiency may be linked to an expansion of B cells in primary mice that is associated with reduced lifespan of KO mice.

Figure 1. KO mice have splenomegaly and increased BCL-6 expression in B cells compared with WT mice; and RhoH re-expression decreases proliferation of KO mouse cells with lymphoma characteristics. (a) Survival curves of WT and KO mice, n = 66 and 65 respectively. **p < 0.01, Log-rank test. (b) Spleen weights of WT and KO mice under 12 months-age, n = 23 and 12, respectively. (c) Photomicrograph of representative sections from spleens of age-matched WT and KO mice stained with haematoxylin and eosin (H&E) (20X magnification); and immunohistochemistry stains with anti-CD19, anti-BCL-6, and anti-Ki-67 antibodies (20X; inset: 100X magnification). Scale bars are 100 µm. (d) Percentage of Ki-67 positive cells in the white pulp areas of WT spleen versus KO spleen as determined by immunohistochemistry (each value is an average of the percentage of Ki-67 positive cells calculated in 5 different areas in each sample; n = 4). (e) In vitro growth of KO CD19⁺ B cell line transduced with RhoH-expressing retrovirus vector (KO-RhoH) versus empty vector (KO-Control), n = 4. Data show mean ± SD. * p < 0.05, *** p < 0.0001.

Rhoh^KOBcl-6^Tg mice show more rapid disease progression in a murine model of lymphoma with increased proliferation of lymphoma cells

To further examine the potential role of RhoH in dysregulated B cell development and lymphomagenesis, we crossed WT and KO mice with the Iµ-HABcl-6 transgenic (Bcl-6^Tg) DLBCL mouse model [17]. In this mouse model, a HA-tagged full length murine BCL6 coding sequence was inserted into the downstream of the immunoglobulin heavy chain (IgH) Iµ promoter on mouse chromosome 12 in murine ES cells. The recombinant locus can produce a chimeric mRNA (Iµ-HABCL6) whose transcription is driven by the Iµ promoter, which is detected in both the germinal centre (GC) and plasma cells of IµHABCL6 mice [17]. The resulting Rhoh^KOBcl-6^Tg (KOTg) and Rhoh^WTBcl-6^Tg (WTTg) mice were observed for disease development. KOTg mice died more quickly than WTTg mice (Figure 2(a), Table 1) and this premature death was associated with development of severe splenomegaly more rapidly (Figure 2(b,c)). Diseased WTTg and KOTg mice displayed loss of splenic architecture and invasion of spleen parenchyma by large, heterogenous lymphoid cells resembling human DLBCL cells (Figure 2(d)). In contrast to WTTg mice, immunohistochemistry revealed that KOTg lymphoma cells had even higher BCL-6 expression and were uniformly and strongly positive for IRF4, suggesting that loss of RhoH alters the phenotype of lymphoma at the cellular level according to Hans criteria [18] to a more aggressive phenotype in the Iµ-HABcl-6 transgenic (Bcl-6^Tg) DLBCL mouse model. We next compared proliferation and apoptosis of WTTg and KOTg lymphoma cells. The percentage of Ki-67-positive cells was significantly increased in KOTg splenic lymphoma cells compared to WTTg lymphoma controls (Figure 2(e)). In contrast, there was no significant difference in the rate of apoptosis as measured by Cleaved Caspase-3 (CC3) (Figure 2(f,g)). Taken together, these results reveal that deletion of Rhoh is associated with a more aggressive disease phenotype in this Bcl-6-driven lymphoma mouse model, with increased proliferation of lymphoma cells.

Table 1. Median survival of WTnTg, WTTg, and KOTg mice

Median survival (Days)
WTnTg	718
WTTg	492
KOTg	359

n = 82, 67, and 69, respectively.

a. p < 0.001, Log-rank test comparing survival for WTnTg mice with WTTg mice.

b. p < 0.001, Log-rank test comparing survival for WTnTg mice with KOTg mice.

c. p < 0.01, Log-rank test comparing survival for WTTg mice with KOTg mice.

Figure 2. KOTg mice show significantly shorter lifespans, increased splenomegaly, and increased lymphoma cell proliferation compared with WTTg mice. (a) Survival curves of WTnTg, WTTg, and KOTg mice, n = 82, 67, and 69 respectively. ** p < 0.01, *** p < 0.001, Log-rank test. (b) Representative photographs of spleens of age-matched WTnTg, WTTg, and KOTg mice. (c) Spleen weights of WTTg and KOTg mice with lymphoma at pre-terminal stage, n = 20 and 25, respectively. (d) Photomicrograph of representative sections from spleens of WTnTg, WTTg, and KOTg mice stained with haematoxylin and eosin (H&E) (40X magnification); and immunohistochemistry stains with anti-BCL-6 and anti-CD138 or anti-IRF4 antibodies (100X magnification). Scale bars are 100 µm. (e) The percentage of Ki-67⁺ CD19⁺ lymphoma cells in WTTg (n = 4) and KOTg (n = 6) spleens as determined by flow cytometry. (f) Number of Cleaved Caspase-3 (CC3)⁺ lymphoma cells derived from WTTg and KOTg spleens as determined by immunohistochemistry, n = 3 and 4, respectively. (g) Photomicrograph of representative sections from spleens of WTTg and KOTg mice immunohistochemistry stained with anti-CC3 antibody. For (c), (e) and (f), data show mean ± SD; student t-test was used for data analysis: * p < 0.05; ns: no significant difference.

RhoH interacts with KAISO and affects the cellular content of KAISO, expression of its target Bcl-6, and downregulation of the BCL-6 target Blimp-1 in lymphoma cells

We have previously reported that RhoH interacts and co-localizes in the nucleus with KAISO, a dual-specific, Broad complex, Tramtrak, Bric-a-brac/Pox virus, Zinc finger (POZ-ZF) transcription factor [19]. In addition, knockdown of RhoH leads to decreased nuclear localization of KAISO in Jurkat T cells [19]. KAISO has also been shown to act as a key regulator of spleen germinal centre (GC) formation by repressing Bcl-6 expression in splenocytes [20]. As KAISO knockout mice develop splenomegaly and large diffuse germinal centres associated with increased expression of Bcl-6 and c-Myc, we hypothesized that the increased expression of BCL-6 seen in KOTg lymphoma cells could be related to decreased nuclear content of KAISO in the absence of RhoH. Immunoblot analysis showed decreased KAISO protein expression in the whole cell lysates of KOTg lymphoma cells when compared with WTTg lymphoma cells (Figure 3(a,b)). Using fractionated lysates isolated from WTTg and KOTg lymphoma cells, we performed immunoblot analysis to examine the expression of KAISO in those fractions. While nuclear KAISO was decreased in KOTg lymphoma cells compared to WTTg lymphoma cells (Figure 3(a,b)), cytoplasmic KAISO was not significantly different in these cells (Figure 3(a,b)). In order to confirm that KAISO and RhoH interact in this lymphoma model, we re-expressed RhoH in B cells derived from KOnTg mice cells using retrovirus-mediated gene transfer. As seen in (Figure 3(c)), HA-tagged RhoH bound to KAISO in RhoH-transduced Rhoh^KO B cells, confirming the interaction of KAISO and RhoH in those cells using co-immunoprecipitation.

Figure 3. RhoH interacts with KAISO and deletion of Rhoh is associated with decreased nuclear content of KAISO in lymphoma cells. (a) Western blot analysis showing the expression of KAISO in the whole cell lysates (top panel), cytoplasmic (middle panel) and nuclear (lower panel) fractions of sorted CD19⁺ spleen cells isolated from WTTg and KOTg mice. (b) Intensity of KAISO Western blot bands by Image J analysis in the whole cell lysates, cytoplasmic, and nuclear fractions of sorted CD19⁺ spleen cells isolated from WTTg and KOTg mice. KAISO band intensity is normalized to the internal control, and those values are normalized to the WTTg KAISO internal control normalized values, n = 3. (c) Western blot analysis of immunoprecipitation of whole cell lysates from a malignant KOnTg B cell line transduced with a retroviral vector expressing HA-RhoH (HA-RhoH) using an anti-HA monoclonal antibody or IgG control followed by immunoblotting with anti-KAISO monoclonal antibody. (d) KAISO protein expression by Western blot analysis in the whole cell lysates (top panel), cytoplasmic (middle panel), and nuclear (lower panel) fractions of a KOTg lymphoma cell line transduced with empty vector (Control) or HA-RhoH (RhoH). (e) Intensity of KAISO Western blot bands by Image J analysis in the whole cell lysates, cytoplasmic, and nuclear fractions of a KOTg lymphoma cell line transduced with empty vector (Control) or HA-RhoH (RhoH). For (a), (b), (d) and (e), β-Actin (ACTB) was used as a loading control for whole cell lysates, α-TUBULIN (TUBA) and LAMIN B1 (LMNB) were used as loading controls for cytoplasmic and nuclear fractions, respectively. KAISO band intensity is normalized to these internal controls, and those values are normalized to the KOTg-Control KAISO internal control normalized values, n = 3. All of the Western blots were performed three times, and representative figures of those blots are shown. Data show mean ± SD, * p < 0.05; ** p < 0.01; *** p < 0.001; ns: no significant difference (f) Subcellular localization of KAISO detected by immunofluorescence staining in a KOTg lymphoma cell line transduced with empty vector (Control) or HA-RhoH (RhoH). Immunofluorescence shows staining with anti-KAISO antibody (red) and nuclear staining with Hoechst33342 (blue). The arrows denote individual cells stained with KAISO. Scale bars are 5 µm. The images were captured by Leica TCS SP8 (STED One) at 100X magnification.

To test whether the re-expression of RhoH can affect the cytoplasmic and nuclear distribution of KAISO, B cell lymphoma cell lines established from enlarged spleens of KOTg mice were transduced with a retrovirus vector expressing HA-tagged RhoH. With immunoblot analysis from the whole cell lysates, we first show that RhoH re-expression increases the total protein expression of KAISO in the KOTg-RhoH cells compared to the KOTg control cells (Figure 3(d,e)). Then, using the isolated fractionated lysates, we show that RhoH re-expression increased the level of KAISO in both the cytoplasm and nucleus of KOTg-RhoH cells compared to KOTg control cells (Figure 3(d,e)). The increased amount of KAISO protein in KOTg-RhoH cells was further confirmed by immunofluorescence staining (Figure 3(f)). Together, these data provide evidence that the interaction between RhoH and KAISO affects the protein levels of the transcription factor KAISO.

As noted above, direct repression of Bcl-6 by KAISO has been reported previously in germinal centre splenocytes, and de-repressed Bcl-6 expression was associated with an increased proliferation of GC cells and splenomegaly in KAISO knockout mice [20]. To further test whether the up-regulation of BCL-6 protein expression observed in KOTg lymphoma cells was due to the release of KAISO-mediated suppression of Bcl-6, we performed a KAISO ChIP assay in KOTg lymphoma cells transduced with RhoH (RhoH). ChIP-qPCR analysis revealed a significant enrichment of KAISO binding at R1 (bp, −900~-450) and R3 (bp, +2165~+2423), two regulatory regions of the Bcl-6 promoter, in the KOTg-RhoH cells compared to KOTg control cells (Figure 4(a)). In contrast, no significant enrichment of the KAISO binding was detected at the Flt3 gene locus in the KOTg-RhoH cells used as a negative control compared to KOTg control cells, suggesting the binding of KAISO at the Bcl-6 promoter was specific. These results strongly indicate that lack of KAISO occupation at the Bcl-6 promoter region may directly de-repress the transcription of the Bcl-6 gene.

Figure 4. KAISO binds to the Bcl-6 promoter, and deletion of Rhoh leads to de-repression of the KAISO target Bcl-6 and the downregulation of the BCL-6 target Blimp-1. (a) Chromatin Immunoprecipitation (ChIP) was performed in a KOTg lymphoma cell line transduced with empty vector (Control) or HA-RhoH (RhoH) against anti-KAISO antibody or IgG control followed by qPCR with primers specific for known KAISO binding sites at the Bcl-6 promoter, including the two regulatory regions R1 (bp, −900~-450) and R3 (bp, +2165~+2423). The Flt3 gene promoter was selected as a negative control. n = 3; data show mean ± SD, * p < 0.05; ns: no significant difference. (b) BCL-6 expression by Western blot analysis in the whole cell lysates (top panel), cytoplasmic (middle panel), and nuclear (lower panel) fractions of sorted CD19⁺ spleen cells harvested from WTnTg, WTTg, and KOTg mice. (c) Intensity of BCL-6 Western blot bands by Image J analysis in the whole cell lysates, cytoplasmic, and nuclear fractions of sorted CD19⁺ spleen cells harvested from WTnTg, WTTg, and KOTg mice. BCL-6 band intensity is normalized to the internal control, and those values are normalized to the WTTg BCL-6 internal control normalized values, n = 4 for the whole lysates and n = 3 for the cytoplasmic and nuclear fractions. (d) BLIMP-1 expression by Western blot analysis in whole cell lysates of sorted CD19⁺ spleen cells isolated from WTnTg, WTTg, and KOTg mice shown. (e) Intensity of BLIMP-1 Western blot bands by Image J analysis in the whole cell lysates of sorted CD19⁺ spleen cells harvested from WTnTg, WTTg, and KOTg mice. BLIMP-1 band intensity is normalized to the internal control, and those values are normalized to the WTTg BLIMP-1 internal control normalized values, n = 4. α-TUBULIN (TUBA) and LAMIN B1 (LMNB) were used as loading controls for cytoplasmic and nuclear fractions, respectively, and β-Actin (ACTB) was used as a loading control for whole cell lysates. All of the Western blots were performed four times for the whole cell lysates and three times for the cytoplasmic and nuclear fractions. Representative figures of those blots are shown. Data show mean ± SD, * p < 0.05; *** p < 0.001; ns: no significant difference.

BCL-6 has been shown to directly repress expression of the transcription factor BLIMP-1 [21,22] and deficiency of BLIMP-1 was demonstrated to interfere with the differentiation of B cells into plasma cells [23]. We conducted immunoblot analysis of whole cell or fractionated cell lysates isolated from primary WTTg and KOTg lymphoma cells to analyse the expression of BCL-6. BCL-6 protein expression was consistently increased especially in the nucleus of sorted CD19⁺ KOTg lymphoma cells compared with WTTg lymphoma controls (Figure 4(b,c)). Immunoblot analysis of BLIMP-1 indicated that the expression of BLIMP-1 in whole cell lysates was consistently reduced in KOTg lymphoma cells compared to normal B cells or WTTg lymphoma cells (Figure 4(d,e)). These data further support that absence of RhoH is associated with decreased expression of Blimp-1, an important gene for the terminal differentiation of B cells to plasma cells in this murine model.

Higher expression of BCL-6 and decreased expression of BLIMP-1 associated with Rhoh deficiency can be reversed by RhoH re-expression

Restoration of RhoH expression in Rhoh-deficient KOTg lymphoma cells using gene transfer led to reduced overall BCL-6 expression (Figure 5(a)) in these cells compared to KOTg controls. BCL-6 expression assessed by anti-BCL-6 antibody in lymphoma cells derived from Iµ-HABcl-6 transgenic mice detects both the transgenic HA tagged-BCL-6 protein and endogenous BCL-6 protein. Since HA-BCL-6 expression detected by anti-HA antibody was similar between KOTg and KOTg-RhoH cells (Figure 5(a,b)) and we had previously demonstrated increased BCL-6 expression in the spleens of KO mice (Figure 1(c)), we surmised that increased expression of BCL-6 in KOTg lymphoma cells was mainly the result of changes in endogenous Bcl-6 gene expression. This was confirmed using B cells derived from KOnTg-RhoH mice, which lack expression of transgenic Bcl-6. Re-expression of RhoH in KOnTg-RhoH B cells was associated with decreased BCL-6 protein levels detected by immunoblot analysis (Figure 5(c,d)) and decreased Bcl-6 mRNA expression by quantitative RT-PCR (qRT-PCR) (Figure 5(e)). Accordingly, the expression of BLIMP-1, a direct target of BCL-6 repression, was increased in KOnTg-RhoH cells compared to KOnTg controls (Figure 5(c,f)), suggesting that the re-expression of RhoH affects not only endogenous Bcl-6 expression but also the expression of genes regulated by BCL-6. Overall, these results indicate that RhoH deficiency leads to deregulated BCL-6 expression that can be partially reversed by the re-expression of RhoH and that in this lymphoma model, RhoH affects BCL-6 transcription through a transcription cascade via RhoH interacting protein KAISO, which is a repressor of BCL-6.

Figure 5. RhoH re-expression in lymphoma cells reverses alterations of BCL-6 and BLIMP-1 expression. (a) Western blot analysis showing the expression of BCL-6, HA tagged BCL-6, or HA tagged RhoH in the whole cell lysates of a KOTg lymphoma cell line transduced with empty vector (Control) or HA tagged RhoH (RhoH). (b) Intensity of total BCL-6 (BCL-6) and exogenous BCL-6 (HA tagged BCL-6) Western blot bands by Image J analysis in the whole cell lysates of a KOTg lymphoma cell line transduced with empty vector (Control) or HA tagged RhoH (RhoH). BCL-6 and HA tagged BCL-6 band intensity is normalized to the internal control, and those values are normalized to the KOTg-Control BCL-6 or HA tagged BCL-6 internal control normalized values, respectively, n = 3. (c) Western blot analysis showing the expression of BCL-6, HA tagged RhoH, and BLIMP-1 in the whole cell lysates (top panel), cytoplasmic (middle panel), and nuclear (lower panel) fractions of a KOnTg B cell line transduced with empty vector (Control) or HA tagged RhoH (RhoH). (d) Intensity of BCL-6 Western blot bands by Image J analysis in the whole cell lysates, cytoplasmic, and nuclear fractions of a KOnTg B cell line transduced with empty vector (Control) or HA tagged RhoH (RhoH). BCL-6 band intensity is normalized to the internal control, and those values are normalized to the KOnTg-Control BCL-6 internal control normalized values, n = 3. (e) Bcl-6 mRNA expression by qPCR in a KOnTg B cell line transduced with empty vector (Control) or RhoH (RhoH) (right panel). β-Actin (Actb) was used as an internal control, n = 6; mean ± SD, * p < 0.05.) (f) Intensity of BLIMP-1 Western blot bands by Image J analysis in the whole cell lysates, cytoplasmic, and nuclear fractions of a KOnTg B cell malignant cell line transduced with empty vector (Control) or HA tagged RhoH (RhoH). BLIMP-1 band intensity is normalized to the internal control, and those values are normalized to the KOnTg-Control BLIMP-1 internal control normalized values, n = 3. For (a) and (c), β-Actin (ACTB) was used as a loading control for whole cell lysates; for (c), α-TUBULIN (TUBA) and LAMIN B1 (LMNB) were additionally used as loading controls for cytoplasmic and nuclear fractions, respectively. All of the Western blots were performed three times, and representative figures of those blots are shown. Data show mean ± SD, * p < 0.05; ** p < 0.01; *** p < 0.001; ns: no significant difference.

Re-expression of RhoH in lymphoma cells reduces proliferation in vitro

To implicate RhoH deficiency in the increased cellular proliferation seen in vivo, we re-expressed RhoH in KOTg lymphoma cells using retrovirus-mediated gene transfer. Re-expression of RhoH in KOTg lymphoma cells led to significantly decreased proliferation measured after 7-days in culture (8.6x10⁶ ± 1.9x10⁵ cells vs 5.9 × 10⁶ ± 1.9x10⁵ cells; KOTg-Control vs KOTg-RhoH, mean ± SD, p < 0.05) (Figure 6(a)). Growth after subsequent days in culture was consistently significantly different as observed through 14 days (Supplemental Figure S2). These results indicate that RhoH regulates proliferation in Bcl-6-driven lymphoma cells. We next determined the contribution of decreased expression of BLIMP-1, which is also altered in Rhoh deficient cells, to increased proliferation of the lymphoma cells. We over-expressed BLIMP-1 in these KOTg lymphoma cells using retrovirus-mediated gene transfer. Over-expression of BLIMP-1 led to significantly decreased cell proliferation measured after 6-days in culture (1.3x10⁶ ± 7.4x10⁴ cells vs 1.8 × 10⁵ ± 1.0x10⁴ cells; KOTg-Control vs KOTg-BLIMP-1, mean ± SD, p < 0.001) (Figure 6(b,c)). Growth was observed to be consistently significantly different through 10 days in culture (Supplemental Figure S2). Overall, our results suggest that the rescue of RhoH expression and the resulting increased expression of the BCL-6 downstream target in this transcriptional cascade, BLIMP-1, may reverse the disease phenotype in KOTg mice.

Figure 6. Re-expression of RhoH in KO lymphoma cells leads to decreased proliferation. (a) In vitro growth of a KOTg lymphoma cell line transduced with a RhoH-expressing retrovirus vector (KOTg-RhoH) versus empty vector (KOTg-Control), n = 4. (b) In vitro growth of a KOTg lymphoma cell line transduced with a BLIMP-1-expressing retrovirus vector with GFP (KOTg-BLIMP-1) versus empty vector with GFP (KOTg-Control). Flow cytometry was used to identify the GFP⁺ cell populations. The percentage of GFP⁺ cells on Days 4 and 6 were normalized to the percentage of GFP⁺ cells detected two days post transduction, n = 3. (c) In vitro growth of a KOTg lymphoma cell line transduced with a BLIMP-1-expressing retrovirus vector with GFP (KOTg-BLIMP-1) versus empty vector with GFP (KOTg-Control), n = 3. Flow cytometry was used to identify the GFP⁺ cell populations. Data show mean ± SD. * p < 0.05, ** p < 0.001. (d) Working model for RhoH involvement in tumour cell proliferation. In WTTg cells without RhoH deficiency, RhoH interacts with KAISO, and KAISO is present in the nucleus where it represses the transcription of Bcl-6. Transcription factor BCL-6 represses the Blimp-1 gene, which codes for the BLIMP-1 transcription factor important in the terminal differentiation of B cells into plasma cells. In KOTg cells with RhoH deficiency, decreased KAISO levels result in de-repression of Bcl-6. BCL-6 expression increases, which is followed by decreased BLIMP-1 expression.

Discussion

RhoH is a haematopoietic-specific, GTPase-deficient member of the Rho (Ras homology) family of small GTPases. RhoH was originally discovered as a translation partner of LACZ/BCL-6 and termed Translocation Three Four (TTF) [3] and also found to be mutated in a number of B cell lymphomas [13]. In humans, loss of function mutations of RHOH have been associated with Epidermodysplasia Verruciformis related to abnormal T lymphocyte phenotype and function [24]. In mice, loss of Rhoh is associated with lack of proper thymic T cell selection and T cell lymphopenia, due to defective proximal T cell receptor signalling, whereas B cell development appears relatively normal [6,25]. While implicated in both immunological function and in various haematologic malignancies, the role of RhoH in tumour development is not clear. For instance, mutations in the RHOH non-coding regions in B cell malignancies may be associated with reduced expression, and decreased expression of RhoH has been shown to predict poorer outcomes in AML, chronic lymphoid leukaemia (CLL) and Hairy Cell Leukaemia [8,26]. In some cases, it has been hypothesized that the pathogenesis is related to the lack of RhoH inhibition of Rac1 and downstream effectors, including phosphoinositol-3-kinase and the Rho guanine exchange factor VAV1, or effects on myeloid differentiation. In contrast, in some instances increased expression of RhoH has been noted in leukaemia cells, specifically in CLL [9], and RhoH deficiency in a murine model of CLL reduced accumulation of leukaemic cells. Thus, the role(s) of RhoH in transformation of haematopoietic cells remains incompletely understood.

In this study, we provide direct evidence that RhoH regulates a transcription pathway modulating the expression of BCL-6 associated with expansion of B cell populations over time in vivo, resulting in a reduced lifespan of Rhoh^−/− (KO) mice. Further, in a model of BCL-6 dependent B cell malignancy, we demonstrate that genetic loss of Rhoh leads to a more rapid disease progression with increased proliferation of lymphoma cells. Mechanistically, the deletion of Rhoh in these lymphoma cells is associated with reduced expression of the RhoH binding partner KAISO, which is associated with the de-repression of Bcl-6, a known KAISO target, and downregulation of the BCL-6 target Blimp-1 (Figure 6(d)). Re-expression of RhoH in Rhoh^KOBcl-6^Tg (KOTg) lymphoma cell lines increases nuclear KAISO levels, decreases BCL-6 protein expression, and leads to reduced proliferation of lymphoma cells. These data suggest that RhoH plays an important role in KAISO mediated transcriptional repression of Bcl-6, a major regulator of B cell development.

We previously demonstrated that RhoH participates in a multi-protein complex with KAISO in Jurkat T cells [19]. KAISO has been defined as a dual-function DNA-binding transcription factor and binds to both methylated CpG dinucleotides in the consensus sequence 5ʹ-CGCG-3ʹ and sequence-specific KAISO binding sites (KBS, 5ʹ-CTGCNA-3ʹ) within target gene promoters through its C-terminal zinc-finger motif [27–29]. Therefore, endogenous KAISO protein levels in the nucleus of mammalian cells is critical for its role in gene regulation [30,31]. Although KAISO has been shown to activate transcription of a subset of target genes [32–34], the predominant role of KAISO is considered to act as a DNA-binding transcriptional repressor. Here, we confirm the interaction of RhoH and KAISO in murine lymphoma cells. The significant enrichment of KAISO binding at the Bcl-6 promoter by ChIP-qPCR shown here and decreased expression of Bcl-6 in KOTg lymphoma cells after RhoH re-expression strongly suggest that KAISO is a transcriptional repressor of Bcl-6 in this model, which is consistent with previous data published about the role of KAISO in normal GC B cells [20]. Additional studies in human lymphoma cells interrogating RhoH and KAISO expression would be informative, particularly as splenomegaly and increased GC cell proliferation due to increased Bcl-6 and c-Myc expression has been noted in the absence of Kaiso in mice [20].

BLIMP-1, a gene whose expression is required for terminal differentiation to plasma cells, is a direct target of BCL-6 [35,36]. Significantly lower BLIMP-1 mRNA expression levels have been found in 26% of an aggressive type of DLBCL harbouring BCL-6 translocations [36]. shRNA-mediated knockdown of BCL-6 expression in the BCL-6-translocated RCK8 and OCI-Ly8 human DLBCL cell lines led to up-regulation of BLIMP-1 mRNA and protein [23]. Consistent with these previous studies, our data show that the reduced expression of BLIMP-1 is associated with increased BCL-6 expression in KOTg lymphoma cells, and re-expression of RhoH leads to both repressed BCL-6 expression and increased BLIMP-1 expression. Thus, our data support the hypothesis that deregulated BCL-6 expression is responsible for suppressing BLIMP-1 expression and provide additional evidence for a role of BLIMP-1 in mediating a more aggressive lymphoma phenotype.

Overall, it appears that dysregulated expression of RhoH alters the function or expression of multiple genes involved in B cell apoptosis and proliferation. Our findings suggest a previously unrecognized role of RhoH in the development of Bcl-6-driven lymphoma via the interaction with KAISO, highlight a role of KAISO in transcriptional silencing of Bcl-6, and suggest that KAISO may function as a tumour suppressor in lymphoma. Since RhoH function may be regulated at the transcriptional level by its unique ITAM-like motif [6] or by lysosomal degradation [37] via chaperone-mediated autophagy, further studies will determine whether modulation of RhoH interaction with KAISO could represent a novel therapeutic approach in B cell malignancies.

Acknowledgments

We appreciate Dr. Laura Pasqualucci for providing us the Iµ-HABcl-6 transgenic mouse model and Dr. Scott Armstrong for helpful discussions. We thank the IDDRC Cellular Imaging Core for immunofluorescence imaging and image processing. We thank Karen Arias, Maria Suarez, Mursal Hassan, Timothy Colby, and Teresa Ortiz for assistance in manuscript formatting and submission.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplementary Material

Supplemental data for this article can be accessed here.

Additional information

Funding

This work was supported by National Institutes of Health grants CA202756, P50CA206963, and CA113969 (to D.A.W.), R01 CA196703 (to R.C), and P50 HD105351 (to the IDDRC Cellular Imaging Core).