Loss of RhoB Expression Enhances the Myelodysplastic Phenotype of Mammalian Diaphanous-Related Formin mDia1 Knockout Mice

Myelodysplastic syndrome (MDS) is characterized by ineffective hematopoiesis and hyperplastic bone marrow. Complete loss or interstitial deletions of the long arm of chromosome 5 occur frequently in MDS. One candidate tumor suppressor on 5q is the mammalian Diaphanous (mDia)-related formin mDia1, encoded by DIAPH1 (5q31.3). mDia-family formins act as effectors for Rho-family small GTP-binding proteins including RhoB, which has also been shown to possess tumor suppressor activity. Mice lacking the Drf1 gene that encodes mDia1 develop age-dependent myelodysplastic features. We crossed mDia1 and RhoB knockout mice to test whether the additional loss of RhoB expression would compound the myelodysplastic phenotype. Drf1 −/− RhoB −/− mice are fertile and develop normally. Relative to age-matched Drf1 −/− RhoB +/− mice, the age of myelodysplasia onset was earlier in Drf1 −/− RhoB −/− animals—including abnormally shaped erythrocytes, splenomegaly, and extramedullary hematopoiesis. In addition, we observed a statistically significant increase in the number of activated monocytes/macrophages in both the spleen and bone marrow of Drf1 −/− RhoB −/− mice relative to Drf1 −/− RhoB +/− mice. These data suggest a role for RhoB-regulated mDia1 in the regulation of hematopoietic progenitor cells.

Introduction mDia-family formins assemble linear actin filaments and modulate microtubule dynamics in response to adhesive and proliferative stimuli [1]. They are regulated by Rho-family small GTP-binding proteins such as RhoB [2,3]. Rho GTPases activate formins through direct binding and disruption of an autoinhibitory mechanism mediated by regulatory domains that flank the actin/ microtubule-binding formin homology-2 (FH2) domain [1]. While the roles of mDia formins in directed cell migration, cell division, and development are well established [4], only recently have genetargeting experiments in mice revealed roles for mDia family formins in immune and myeloid cell proliferation [5,6,7].
The DIAPH1 gene encoding human mDia1 (located at 5q31.3) lies between the two commonly deleted regions mapped by conventional cytogenetics in myelodysplastic syndrome (MDS) patient samples [8]. 5q-MDS is characterized by peripheral cytopenias and ineffective hematopoiesis [9]. How defects in one or more 5q genes trigger MDS or contribute to malignant progression in conjunction with additional chromosomal abnormalities remains unknown [10]. Gene targeting experiments on the murine DIAPH1 homolog Drf1 show that loss of mDia1 expression impairs the growth control of hematopoietic progenitors [6]. However, the specific mechanism by which loss of mDia1 expression triggers the MDS-like phenotype is currently under investigation.
The small GTPase RhoB binds to both mDia1 and mDia2 on endosomes and has a role in endosome trafficking [2,3]. RhoB also plays an important role in the apoptotic response to DNA damage [11], and loss of RhoB expression has been shown to correlate with late-stage malignancy [12,13]. Mice lacking RhoB alone do not show any signs of myelodysplasia or any developmental or fertility defects, but Ras-transformed mouse embryonic fibroblasts (MEFs) from these mice are resistant to apoptosis in the presence of farnesyltransferase inhibitors (FTIs), doxorubicin, or taxol [14,15,16]. Together, these studies suggest RhoB possesses tumor suppressor activity.
In this study, we hypothesized that the myelodysplasia observed in Drf1-null mice would be enhanced by the additional loss of one of its regulators, RhoB. After examination of the peripheral blood, bone marrow, and spleen hematopoietic progenitor cells, we show that Drf1 2/2 RhoB 2/2 mice develop age-dependent myelodysplasia before Drf1 2/2 RhoB +/2 mice. These results are consistent with a model for disease progression in MDS that includes the alteration of multiple tumor suppressors in hematopoietic stem or progenitor cells.

Gene targeting
The gene targeting and genotyping of Drf1-null mice were exactly as previously described [17]; targeting and genotyping of RhoB-null mice were as described in [15]. The mice used in this study were a mixed 129/B6 genetic background. For all experiments, data was acquired from six Drf1 2/2 RhoB +/2 and six Drf1 2/2 RhoB 2/2 mice at 100 days of age, and from eight Drf1 2/2 RhoB +/2 and nine Drf1 2/2 RhoB 2/2 mice at 400 days of age.

Ethics Statement
All experiments performed were approved in advance by the Van Andel Research Institute Institutional Animal Care and Use Committee.

Flow cytometry analysis
Peripheral blood, bone marrow, and splenic single-cell suspensions were characterized by flow cytometric analysis. Peripheral blood was extracted from the heart using a syringe equipped with a fine gauge needle. Bone marrow was flushed from femurs using a syringe with a fine gauge needle and 3 mL of PBS. Single-cell suspensions of the spleen were obtained by mincing tissue with glass slides and subsequent passage and scraping of tissue in a ThermoShandon biopsy bag (Thermo Fisher Scientific).
Cells were incubated for 15 min at 20uC in the dark. Incubation was followed by addition of 16 FACSLyse reagent (Becton Dickinson) for 15 min at 20uC in the dark. After RBC lysis, the remaining cells were washed in 2 mL PBS with 0.1% sodium azide. Cells were fixed in 1.0% methanol-free formaldehyde (Polysciences, Inc.) in PBS containing 0.1% bovine serum albumin and refrigerated at 4.0uC until acquisition. Appropriate subclass and negative controls were used to detect nonspecific binding of antibody and autofluorescence. A minimum of 10,000 events for fresh mononuclear cells and 5,000 events for splenic cells were acquired. Flow cytometric analyses were conducted using either a FACSCalibur 4-color or a FACSAria 12-color flow cytometer (Becton Dickinson). Data were analyzed using Becton Dickinson CellQuest Pro and FACSDiVa software.

Monoclonal antibodies
The following monoclonal antibodies were used: anti-

Cell cycle analysis, complete blood count (CBC), and statistics
Cell cycle analyses used propidium iodide (Sigma) in a modified Vindelov's preparation. A minimum of 10,000 events were collected by flow cytometry. Data were analyzed using Becton Dickinson CellQUEST Pro and Verity House ModFIT LT software. CBC analysis was performed using a VetScan HM2 Hematology System (Abaxis). All statistical analysis was performed using the one-tailed Mann Whitney test for significance. Each point on scatter plots represent a single mouse, and each plot includes a horizontal line to indicate the median of the data. Scatter plots and statistics were performed using GraphPad Prism 5.0 software.

Results
We previously reported that Drf1 2/2 and Drf1 +/2 mice develop age-dependent myelodysplasia, typically around 450 days of age [6]. We asked whether the additional homozygous loss of RhoB would enhance disease progression relative to Drf1-null mice that still contain a functional allele of RhoB. Previous reports have shown that RhoB +/2 cells are indistinguishable from RhoB +/+ cells in their apoptotic response to DNA damage [16]. We characterized multiple mice of each genotype at 100 and 400 days of age to assess the presence of myelodysplastic features.
Based on our previous work in Drf1-null mice [6] and the recent finding that the mDia2 formin contributes to erythroblast enucleation [18], we examined the peripheral blood to determine if there was morphological evidence of erythrocyte dysplasia. Peripheral blood smears from Drf1 2/2 RhoB 2/2 mice showed a marked elevation in abnormally shaped erythrocytes compared with blood from Drf1 2/2 RhoB +/2 mice. The erythrocytes were often spiked (echinocyte) or teardrop in appearance ( Fig. 2A), consistent with dysplastic features observed in patients with MDS. Drf1 2/2 RhoB +/2 mice did show signs of dysplasia at 400 days, with several teardrop-shaped erythrocytes, but to a lesser extent than Drf1 2/2 RhoB 2/2 mice. CBC analysis of peripheral blood revealed a significant increase in the WBC count of Drf1 2/2 RhoB 2/2 mice relative to that of Drf1 2/2 RhoB +/2 mice at 400 days of age (Fig. 2B). Abnormal platelet counts are sometimes observed in certain subsets of MDS, but we did not observe any significant differences in these mice (Fig. 2C).
We then focused our analysis specifically on the bone marrow and spleen to examine myelopoiesis in these compartments. Flow cytometry was used to determine the percentage of lymphocytes, monocytes, and granulocytes based on the pan-leukocyte marker CD45. Using the gating strategy outlined previously [6], we found that 400-day-old Drf1 2/2 RhoB 2/2 mice had an increased percentage of granulocytes and a concomitant decrease in lymphocytes in the bone marrow compartment (Fig. 3A). Analysis of splenic single-cell suspensions isolated from 400-day-old mice showed a similar increase of granulocytes in Drf1 2/2 RhoB 2/2 mice (Fig. 3B).
To further examine potential differences in myelopoiesis between Drf1 2/2 RhoB +/2 mice and Drf1 2/2 RhoB 2/2 mice, we performed flow cytometry on cells from the bone marrow and spleen to detect levels of F4/80 (a pan macrophage marker) and CD11b (integrin aM; monocyte development marker). By 400 days of age, Drf1 2/2 RhoB 2/2 mice had an increased percentage of F4/80 + cells in the spleen, but we observed no significant difference in the bone marrow (Fig. 3C). On the other hand, CD11b + cells were significantly elevated in both the marrow and spleen by 400 days of age. We also examined cellular expression of the marker CD29 (b1 integrin; important for homing and retention in lymphoid organs). We found an increased percentage of CD29 + cells in the marrow, with an even more pronounced increase in the spleen of Drf1 2/2 RhoB 2/2 mice relative to Drf1 2/2 RhoB +/2 mice (Fig. 3C). These results are consistent with the observed increase in myeloid cell proportion found by histopathology (Fig. 1).
Finally, we analyzed the levels of TER-119 (erythroid-specific marker) and CD71 (transferrin receptor; a marker of proliferating erythroid precursors) in the marrow and spleen of Drf1 2/2 RhoB 2/2 and Drf1 2/2 RhoB +/2 mice. While there were no differences in the bone marrow, the levels of CD71 + and TER-119 + cells were elevated in the spleens of 400-day-old Drf1 2/2 RhoB 2/2 mice (Fig. 4, A and B). An increase in erythroid precursors was also evident in H&E-stained splenic sections (data not shown). Consistent with this observation and the presence of splenic EMH, we also found a substantial increase in splenic cells undergoing S phase in 400-day-old Drf1 2/2 RhoB 2/2 mice relative to the number in Drf1 2/2 RhoB +/2 mice (Fig. 4C).

Discussion
MDS is thought to arise because of multiple alterations in a hematopoietic stem cell [10]. We previously found that knocking out mDia1 expression in mice leads to the age-dependent development of myelodysplasia [6]. Here, we show that the additional knockout of RhoB expression in Drf1-null mice accelerates the progression to myelodysplasia. Several candidate tumor suppressor genes in humans reside on the long arm of chromosome 5 [8]. One of these genes is DIAPH1, which encodes the actin assembly protein mDia1. While genetic ablation of mDia1 expression alone in mice resembles 5q-MDS, it is likely that mDia1 acts in concert with multiple other genes on the same commonly deleted region to suppress malignancy [8].
MDS can progress to a more advanced stage and ultimately develop into leukemia. The transformation events involved in disease progression are thought to include genes that contribute to cell-cycle control, apoptosis, and differentiation [10]. The small GTPase RhoB is required for apoptosis in response to DNAdamaging agents and farnesyltransferase inhibitors [14,16]. Previous reports have shown that RhoB negatively regulates Akt survival signaling and mediates apoptosis in a p53-dependent manner [19,20]. Given its role in apoptosis and its known interactions with the mDia formins, we hypothesized that deletion of RhoB would enhance the progression of MDS in mice lacking mDia1 expression. Our results support this hypothesis and further substantiate our mouse model of age-associated myelodysplasia. It is also interesting to speculate that loss of RhoB expression may interfere with hematopoietic stem cell maintenance, since recent work has highlighted a role for RhoB in stem cell self-renewal [21]. Together, these findings suggest that multiple mechanisms may contribute to the myelodysplastic phenotype in our mice.
Our mouse model uniquely positions us to test some important questions related to the in vivo role of RhoB in the anti-neoplastic affects of FTIs. FTIs have been tested clinically to treat various myelodysplasias, with varying degrees of efficacy [22]. But whether FTIs can alleviate the disease phenotype in Drf1-null or heterozygous mice and whether RhoB is required in vivo to mediate this response remains to be determined. Our results point to the down-regulation of RhoB as a potential marker for late-stage MDS, similar to its diminished expression in other late-stage cancers [12,13]. Examination of the gene expression profile of RhoB in human patient samples that have a more advanced disease signature could help determine better treatment options for patients diagnosed with MDS.
In summary, we show that mice lacking mDia1 and RhoB expression progress to MDS faster than mice lacking mDia1 alone. These data parallel observations of MDS in humans, in which multiple alterations in hematopoietic stem cells contribute to disease pathogenesis. Our mouse model will be useful in characterizing the mechanism of disease progression further and in testing potential therapeutics to treat this chronic disease.