Mice Deficient in Rap1A GTPase Display Reduced Adhesion of Hematopoietic and Mesenchymal Stromal Cells but Intact Hematopoietic Repopulation Potential

Background: The Guanine Triphosphatase (GTPase) Rap1A has been implicated in cell proliferation and migration in different tissues. Its role in the production, adhesion and migration of bone marrow-derived hematopoietic cells is however incompletely understood. Methods: We analysed mice lacking one or both Rap1A alleles and measured their adhesion function and their hematopoietic repopulation capacity using competitive bone marrow transplantation in the CD45.1/CD45.2 transplant model. We further employed flow cytometry and shear stress-dependent flow chamber adhesion assays for phenotypic characterization and analysis of adhesion functions. Results: Rap1A-/mice showed decreased numbers of lymphocytes and hematopoietic colony forming progenitors in the blood, while the numbers of granulocytes, monocytes or platelets were not significantly altered compared to controls. Chemokine CXCL12-induced adhesion of lineage marker depleted (Lin-) bone marrow cells to Vascular Cell Adhesion Molecule (VCAM)-1 and to Intercellular Adhesion Molecule (ICAM)-1 was significantly reduced in Rap1A-/mice. However, competitive repopulation by Rap1A deficient bone marrow cells was not inferior to wild type bone marrow 9 months after transplantation. No outgrowth of Mesenchymal Stromal Cells (MSCs) was observed from bone marrow of Rap1A-/mice, whereas Rap1A+/MSCs displayed reduced expression of adhesion molecules involved in progenitor homing and impaired adhesion to primary endothelial cells and VCAM-1 under shear stress in vitro, as compared to their wild-type counterparts. Conclusions: Rap1A regulates adhesion of immature hematopoietic cells and mesenchymal stromal cells, but appears to be redundant for hematopoietic engraftment. Mice Deficient in Rap1A GTPase Display Reduced Adhesion of Hematopoietic and Mesenchymal Stromal Cells but Intact Hematopoietic Repopulation Potential Erika Fleck1, Stephan Göttig1, Harald Kropshofer2, Juliane Steinmann3, Annette Dorn3, Brigitte Rüster1, Erhard Seifried1, Jürgen Scheele3 and Reinhard Henschler1,4* 1Institute for Transfusion Medicine and Immune Hematology, German Red Cross Blood Service, Medical Center of the Johann Wolfgang Goethe University, Frankfurt (Main), Germany 2F Hoffmann-La Roche Ltd, Pharmaceuticals Division, Product Development, Basel, Switzerland 3Department of Hematology and Oncology, University Hospital Medical Center, Freiburg, Germany 4Blood Centers Zürich and Chur, Swiss Red Cross Blood Service, Zürich and Chur, Switzerland *Corresponding author: Reinhard Henschler, MD, Ph.D., Institute for Transfusion Medicine and Immune Hematology, German Red Cross Blood Service, Medical Center of the Johann Wolfgang Goethe University, Sandhofstraße 1 D 60528 Frankfurt (Main), Germany; Blood Center Zürich SRK, Rütistrasse 19, CH-8952 Zürich-Schlieren, Switzerland, Tel: +41 58 272 5117; E-mail: rhenschler@gmx.de Received July 27, 2016; Accepted August 16, 2016; Published August 22, 2016 Citation: Fleck E, Göttig S, Kropshofer H, Steinmann J, Dorn A, et al. (2016) Mice Deficient in Rap1A GTPase Display Reduced Adhesion of Hematopoietic and Mesenchymal Stromal Cells but Intact Hematopoietic Repopulation Potential. J Stem Cell Res Ther 6: 356. doi: 10.4172/2157-7633.1000356 Copyright: © 2016 Fleck E, et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.


Introduction
The GTPase Ras-associated protein (Rap) 1 belongs to the family of small GTPases, and has been described as a Ras antagonist in oncogenic transformed fibroblasts [1]. The two isoforms of Rap1, Rap1A and Rap1B are known to be implicated in the regulation of immune responses mainly through affecting adhesive functions and activation of lymphocyte populations [2][3][4]. More recently, Rap1B deficiency has been linked to defects in B cell development and angiogenesis [5][6][7]. Moreover, Li et al. [8] reported that neutrophilic granulocytes show a reduced oxidative burst formation in the absence of Rap1A and Jang et al. described a role for Rap1 in prostaglandin E2-induced proliferation of cord blood derived MSCs [9].
However, data are lacking which describe the function of Rap1A in undifferentiated hematopoietic cells or bone marrow-derived MSCs. We therefore investigated whether Rap1A deficient mice show alterations in levels of hematopoietic cells in blood as well as in proliferation and differentiation after bone marrow transplantation. In addition, we explored the phenotype and adhesive function of MSCs isolated from bone marrow of Rap1A deficient mice. We report here that Rap1A deficiency results in lower numbers of lymphocytes and hematopoietic colony forming progenitor cells in blood. Furthermore, Lin-enriched undifferentiated progenitor cells have a reduced capacity to adhere to VCAM-1 under shear stress after chemokine stimulation.
Likewise, Rap1A was essential for establishment of MSCs from bone marrow, Rap1A deficient MSCs displayed lower numbers of surface adhesion molecules and impaired adhesion to endothelial cells and immobilized VCAM-1 under shear stress. However, bone marrow cells from Rap1A -/mice displayed unperturbed engraftment potential compared with wild-type cells.

Isolation and expansion of MSCs
Bone marrow cells were collected from femora of 8-12-week-old mice and plated at a density of 5×10 5 cells/cm 2 in DMEM medium supplemented with 10% FCS, 100 IU/ml penicillin, 0.1 mg/ml streptomycin and 2 ng/ml fibroblast growth factor (PAA Laboratories, Cölbe, Germany). Cells were incubated in a humidified environment at 37°C and 5% CO 2 . Non-adherent cells were discarded after 5 days of culture and the whole medium was replaced by fresh medium. Afterwards, the medium was replaced every 3-4 days within a period of 3-4 weeks. When the layer of adherent spindle-shaped cells became sub-confluent (70-75%), the cells were passaged by trypsin treatment, and the expression of the surface markers CD45, CD105 and CD73 was analyzed by flow-cytometry. As soon as >98% cells displayed the phenotype CD45 -CD105 + CD73 + , their differentiation potential was confirmed by successful induction of osteoplastic and adipogenic differentiation upon addition of NH OsteoDiff Medium or NH AdipoDiff Medium (Miltenyi Biotech, Bergisch-Gladbach, Germany), respectively.

Parallel plate flow chamber adhesion studies
To study interaction with recombinant adhesion ligands, parallel plate flow chambers (µ slides, Ibidi, Munich, Germany) were pre-coated with 20 µg/ml (in phosphate-buffered saline (PBS)/ 1% (v/v) bovine serum albumin (BSA) recombinant murine VCAM-1-immunoglobulin (Ig) or ICAM-1-Ig fusion proteins (Becton Dickinson) for 30 min at room temperature. Chambers were blocked with PBS/1% BSA to avoid unspecific binding 30 min prior to the experiment. Ten minutes before the start of the experiment, chemokine CXCL12 (final concentration 1000 ng/ml, R&D Systems) was immobilized on the bottom of the flow chamber. One hundred thousand Lin-cells were seeded in Hank's buffered salt solution (HBSS; Gibco, Karlsruhe, Germany) supplemented with 0.1% BSA into the flow chamber per channel and allowed to adhere for 3 min. Then, medium flow was started with 37°C prewarmed HBSS/0.1% BSA using a precision pump (Hugo Sachs Elektronik Harvard Apparatus GmbH, March, Germany) at a calculated shear stress of 0.35 dynes/cm 2 . Shear stress was increased stepwise every 30 s. Numbers of adherent cells were determined using a CCD camera and counted manually.
Interaction of MSCs with endothelial cells or VCAM-1 under shear stress. The coating of the flow chamber basement with endothelial cells was performed 1-2 days prior to the experiment. To this end, 1×10 6 human umbilical vein endothelial cells (HUVECs) were cultured in the flow chamber slide pre-coated with 0.1% gelatin at 37°C in a humidified atmosphere with 5% CO 2 . Experiments were performed after overnight stimulation with 10 ng/ml Tumor Necrosis-Factor (TNF)-α (R&D Systems).
To study interactions of MSCs with HUVECs at low shear stress, 1×10 6 wild-type or Rap1A +/-MSCs were flushed in pre-warmed (37°C) Hank´s Balanced Salt Solution (HBSS) supplemented with 0.5% BSA over a confluent HUVEC monolayer at a calculated shear stress of 0.1 dynes/cm 2 , and numbers of interacting cells were determined microscopically. To study firm adhesion of MSCs, 1×10 6 MSCs were added in 100 µl 37°C pre-warmed HBSS to the chamber and allowed to adhere to HUVECs for 10 min with or without previous TNF-α stimulation. Subsequently, pre-warmed HBSS with 0.5% BSA was pumped over the cells under a constant shear stress of 0.1 dynes/cm 2 until non-adherent cells were flushed away. Cells were enumerated and, subsequently, the shear stress was increased to 2 dynes/cm 2 . After 10 min, the remaining adherent MSCs were counted.
The adhesion to VCAM-1 was studied after coating the flow chamber for 30 min at ambient temperature with VCAM-1 (30 µL; 20 µg/ml). Unspecific binding was blocked by adding 50 µl PBS/1% BSA for 10 min. CXCL12 was added at 1000 ng/ml in 100 µl PBS/1% BSA. Ten to the sixth MSCs were added in 100 µl pre-warmed HBSS and left for 5 min before flow was started. Shear stress was initiated at 0.1 dynes/ cm 2 for 30s, and then increased stepwise every 30s reaching up to 2.5 dynes/cm 2 . Numbers of adherent cells were documented at the end of the 30s period for each shear stress level. For all assays, representative pictures of each chamber were taken at each step of the experiment to assess numbers of adherent MSCs.

Statistical analysis
A two-tailed Student's t test was performed to evaluate the significance of the results. P values ≤ 0.05 were considered as significant.

Peripheral blood cell numbers in Rap1A deficient mice
To addressed the potential role of the GTPase Rap1A in hematopoietic cell development, we first determined the levels of mature cell types in blood. Stem cell potential was assessed by competitive hematopoietic repopulation of Rap1A deficient and wild-type bone marrow cells in pre-irradiated mice by bone marrow transplantation. In addition, we asked whether Rap1A deficient undifferentiated hematopoietic cells may be defective in adhesion functions. To confirm the genotype of the knockout mice, PCR amplification of normal and deleted Rap1 alleles with different primer pairs was used as previously published [3] ( Figure 1A). Some genotypings yielded weak bands in the heterozygous animals. From a total of 97 littermates, we obtained 52 (53.6%) heterozygous animals, corresponding to the expected Mendelian inheritance ratio. In contrast, we obtained only 8 (8.2%) Rap1A -/mice. The numbers of neutrophils, monocytes, and platelets were not significantly different in the blood of wild-type or knockout mice ( Figure 1B). However, we found a significant decrease in the number of lymphocytes compared with wild-type controls ( Figure  1B). In addition, the frequency of granulocyte-macrophage myeloid CFCs dropped threefold in blood of Rap1A -/mice compared to controls ( Figure 1C). These data indicate an influence of Rap1A on levels of lymphocytes and some hematopoietic progenitor cells in blood of the mice.

Rap1A is implicated in adhesion of hematopoietic progenitors
We next explored whether the reduced numbers of hematopoietic progenitor cells found in the blood of Rap1A -/mice were associated with a decreased expression of relevant adhesion molecules, mediatinge interaction of hematopoietic progenitors with endothelial cells. The integrins VLA-4, LFA-1 and Mac-1, which are ligands to VCAM-1 and ICAM-1 on endothelial cells, as well as the CXCL12 receptor CXCR4 and CD44 were generally expressed at similar levels both in wild-type and Rap1A deficient Lin-cells (Figures 2A and 2B). We then analyzed the adhesion function of Lin-progenitors from bone marrow of wild-type, heterozygous or knockout mice to the integrin ligands VCAM-1 or ICAM-1 in a parallel flow chamber. Since the chemokine CXCL12 has been previously implicated in mobilization of hematopoietic progenitor cells and strengthening of integrin-mediated adhesion, we included investigation of the role of co-immobilized CXCL12 on integrin-mediated adhesion of Lin-cells. Already at low (0.35 dynes/cm 2 ) shear stress, and pertaining to high shear stresses of up to 15 dynes/cm 2 , Rap1A -/-Lin-cells showed a slightly decreased adhesion to ICAM-1 but not VCAM-1 compared to wild-type controls ( Figure 2C). However, the CXCL12-induced increase in adhesion to both, VCAM-1 and ICAM-1 was significantly lower in Rap1A deficient Lin-cells compared to wild type controls ( Figure 2C). Thus, although the integrin receptors for VCAM-1 and ICAM-1 as well as CXCR4 were expressed at comparable levels on both wild-type and knockout bone marrow cells, Rap1A deficient Lin-bone marrow cells showed a deficiency in integrin-mediated adhesion stimulated by the chemokine CXCL12.

Role of Rap1A in engraftment after bone marrow transplantation
To investigate whether this adhesion defect also translates into an engraftment defect, we transplanted freshly isolated bone marrow cells from Rap1A -/or Rap1A +/mice together with an equal number of wild-type CD45.1 competitor bone marrow cells into lethally preirradiated recipient mice. As shown in Figure 3A, the proportions of donor type hematopoietic cells were generally similar in bone marrow, blood and spleen 9 months after competitive repopulation of mice with wild-type, Rap1A +/or Rap1A -/bone marrow cells and Ly 5.1 wild-type bone marrow cells. A relatively small 22%) increase in the proportion of donor-derived cells was detected only in the bone marrow of mice receiving Rap1A -/donor cells ( Figure 3A); this difference was however not found in blood or spleen of the same recipient mice. Similar findings were obtained when analyzing subfractions of leukocytes in hematopoietic tissues or blood of the recipient mice ( Figure 3B). Together, these data demonstrate that bone marrow cells from Rap1A deficient mice display adhesion deficits to stromal ligands, but do not harbour overt deficits in competitive hematopoietic reconstitution ability.

Role of Rap1A in establishment of MSCs in culture, and adherence of MSCs to endothelial cells and VCAM-1
We next asked whether bone-marrow derived stromal cells, which are known to be critically implied in the regulation of hematopoiesis, and which are typically represented by MSCs isolated from bone marrow, may be affected by Rap1A deficiency. To this end, whole bone marrow cells from Rap1A +/+ , Rap1A +/and Rap1A -/mice were depleted of hematopoietic cells by plastic adherence and cultured under conditions favoring MSC growth. While we were able to successfully isolate and expand MSCs from Rap1A +/+ and Rap1A +/mice, no MSC outgrowth was seen with bone marrow cells isolated from Rap1A -/mice ( Figure 4A). Still, MSCs could be established from Rap1A +/bone marrow. However, we observed slower growth within the initial 50 days, and an acceleration thereafter in MSC cultures derived from Rap1A +/-   integrin, CD44 and CD62L on Rap1A +/-MSC compared to MSC from wild-type mice by flow-cytometry ( Figure 4B).
To examine whether the differential expression of adhesion molecules translates into differential interaction with endothelial cells, we tested the interaction of MSCs with immobilized HUVEC cells in the parallel plate flow chamber model. In this model, MSCs were studied for their adhesion capacity on TNF-α-pre-stimulated or unstimulated HUVEC at a relatively low shear shear stress of 0.1 dynes/ cm 2 or 2 dynes/ cm 2 , to mimic the conditions found in post-capillary venules. We found a trend, although no statistical significance, towards a lower number of Rap1A +/-MSCs interacting with HUVECs mice compared to wild-type mice ( Figure 4A). These data implicate an adhesion defect of MSCs lacking Rap1A. Immunophenotyping using flow-cytometry revealed that the MSC cultures were positive for MSC-specific surface markers, such as CD73, CD105 or CD13, without notable differences between wild-type and Rap1A +/-MSCs (data not shown). Furthermore, wild-type and Rap1A +/-MSC were equally able to differentiate into adipogenic or osteogenic cells (data not shown).
In search for an underlying cause of the adhesion deficits in Rap1A deficient MSC, we next examined the expression of several adhesion molecules on wild-type and Rap1A +/-MSCs. We found a slightly lower expression of adhesion molecules CD11a, CD11b, α 4 β 7 integrin, β 1 in comparison with wild-type MSCs, independent of whether the HUVEC layer was stimulated with TNF-α or not ( Figure 5A). We next analysed the ability of MSCs to firmly adhere to HUVECs ( Figure  5B). Whereas no significant differences were detected in unstimulated HUVECs, the adhesion capacity of wild-type MSC was significantly higher compared with Rap1A +/-MSC after pre-stimulation of HUVEC with TNF-α ( Figure 5B). Finally, we analysed the adhesion capacity of MSC to immobilized recombinant VCAM-1 ( Figure 5C). Rap1A +/-MSCs were inferior to wild-type MSCs in adhesion to VCAM-1, both in the absence and presence of the stimulator CXCL12. Together, these data demonstrate a role of Rap1A in adhesion to plastic during the outgrowth phase from bone marrow, and in firm adhesion of MSCs via the integrin VLA-4 -VCAM axis and co-stimulated through CXCL12.

Discussion
So far, limited data are available describing the function of Rap1A after genetic ablation in hematopoietic cells. The complete ablation of the Rap1B gene has been shown to lead to embryonic lethality, due to developmental abnormalities during early vasculogenesis, pointing out an important role of Rap1B in mesodermal development [3,5]. It is not clear whether such defects would also contribute to the lower birth rate in the Rap1A deficient mice which we found. Yan et al. [10] have recently reported that a knockdown of either Rap1A or B in endothelial cells by siRNA leads to a loss of vascular permeability and endothelial tube formation. In previous collaborative work with Chavakis et al. [6] we have shown that Rap1A or Rap1B deficiency is associated with a reduction of integrin activation in endothelial cells, using both siRNA mediated knockdown of Rap1A and B, and the Rap1A knockout mouse model. In view of the role of Rap1A and B in endothelial cell function, it was of interest to follow potential defects in hematopoietic cells with regard to adhesion to stromal ligands and hematopoietic engraftment, which are both also known to be regulated by integrins.
Our findings on blood cell numbers of myeloid hematopoietic cells are comparable to our previous results analysing Rap1A deficient T lymphocytes, where we observed a reduction in numbers of circulating T lymphocytes [3]. In addition to the adhesion defect of T lymphocytes on immobilized integrin ligands described earlier [3], we here report a significant adhesion defect to immobilized integrin ligands also in immature myeloid cells of the lin-fraction. So far, an adhesion deficit of hematopoietic cells has only been reported in the human leukemic hematopoietic cell line UT-7 after overexpression of RapGAP1, resulting in downregulation of Rap1 activity [11,12]. Our data with Rap1A deletion in primary cells demonstrates the direct involvement of Rap1A in integrin activation in the progenitors. In addition, we show that primarily the CXCL12-induced adhesion is affected. Other than number of interacting cells previous studies, we have employed defined shear stress conditions. Other authors, using Rap1A knockout mice, have reported a migration defect of splenic B and T lymphocytes in transwell migration assays and in adhesion assays involving the vitronectin and fibronectin analogue CH-296 [8]. They also analyzed transwell migration of Lin-Sca1+ hematopoietic progenitor cells from bone marrow, and found their migration suppressed in the absence of Rap1A. Ghandour et al. [13], found no significant role of the Rap1 activator CalDAG-GEFI in VCAM- A phenotype of Rap1B deficiency in B cell development was reported by Chen et al. [7], who described delayed development of B cells resulting in altered B and T lymphocyte homeostasis in the absence of Rap1B. Awasthi et al. confirmed the finding of a hematopoietic maturation defect of B and NK cells in the absence of Rap1B, implying a role mostly for Rap1B in functional activation and in homing behaviour of NK cells, and a role for the Rac activation regulator IQGAP in this process [12]. Our studies, both in untransplanted Rap1A -/mice and in chimeric mice lacking Rap1A in hematopoietic cells after bone marrow transplantation, do not indicate a major role of Rap1A in B cell maturation. However, a compensation of Rap1A deficiency by increased expression of endogenous Rap1B cannot be excluded also in our model. Yet, we could not analyse Rap1A or B activation separately due to lack of appropriate antibodies.
Since MSCs which have been found to support hematopoiesis show themselves a distinct adhesion and migration behaviour [18], we analysed to what extent Rap1A influences adhesion of bone marrowderived MSCs. We found a reduction by 20-80% of surface integrins on MSCs from Rap1A +/mice compared to their wild-type counterparts ( Figure 4). Consistent with this finding, Rap1A +/-MSCs were inferior to wild-type MSC in the flow chamber model under shear stress, since their adhesion to TNFα-stimulated inflammatory HUVEC cells and also to immobilized VCAM-1 was impaired ( Figure 5). These findings may also explain why MSCs carrying the homozygous deletion of Rap1A were unable to survive under our cell culture conditions. This study, to our knowledge, is the first to study MSCs with genetic ablation of Rap1A. Thankamony and Sackstein [19] have used a Rap1 inhibitor and found a decreased adherence of human MSCs under flow. siRNA mediated inhibition of Rap1A was studied by Jang et al. [9], who found a decreased AKT phosphorylation in MSCs that was associated with a lower proliferation rate under Rap1 inhibition, and by Chen et al. [20] in placental derived MSCs, also after siRNA mediated Rap1 inhibition. In contrast to these studies which employ siRNA or chemical inhibition, both of which cannot 100% selectively target the Rap1A gene alone, our study using genetic ablation shows that Rap1A is indispensable for proliferation initiation of MSCs. A phenotype of defective activation of integrins in Rap1A knockout mice has been described also in endothelial cells [16] and in T cells, resulting in reduced contact dermatitis [21].
Histone methyltransferase G9a is essential for the T cell differention during murine intestinal inflammation [22]. A recent observation has shown that G9a is also involved in T cell adhesion via interaction with lamin B1 [23]. The impairment of adhesion in the Rap1A deficient immature hematopoietic cells and mesenchymal stromal cells might directly caused by prevention of the function of G9a. Since G9a plays a major role in the establishment and maintenance of H3K9me2 [24] and also associates with maintenance of DNA methylation [25], it is interesting to investigate the patterns of H3K9me2 and DNA methylatIon in the Rap1A deficient HSCs/MSCs.
Taken together, our data show a redundant role of Rap1A in the homeostasis of mature lymphocytes, myeloid cells and hematopoietic repopulating cells, but a significant role in the proliferation and adhesion in MSCs/stromal cells. The complete role of both Rap1 allelic products in these processes may be determined as soon as inducible double knockout mouse models become available, e.g., a partial knockdown of both Rap1 alleles or by breeding one or two floxed Rap1B alleles into Rap1A deficient mice and subsequently inducing hematopoietic-specific deletion. This will further clarify the role of this important regulator of integrin activation in hematopoietic and mesenchymal stem cell biology. Moreover, Rap1A may serve as a molecular target for hematopoietic and mesenchymal cell adhesion and/or mobilization processes.

Authors´ Contributions
EF and SG carried out the mouse breeding and genetic analysis, EF and BR performed mouse experimentation and transplantation experiments, cell isolation, cell culture, and the flow cytometric analysis. RH, JS and ES conceived the study, and participated in its design and coordination. EF, HK, JS and AD drafted the manuscript. All authors read and approved the final manuscript.