Strain-Dependent Differences in Bone Development, Myeloid Hyperplasia, Morbidity and Mortality in Ptpn2-Deficient Mice

Single nucleotide polymorphisms in the gene encoding the protein tyrosine phosphatase TCPTP (encoded by PTPN2) have been linked with the development of autoimmunity. Here we have used Cre/LoxP recombination to generate Ptpn2ex2−/ex2− mice with a global deficiency in TCPTP on a C57BL/6 background and compared the phenotype of these mice to Ptpn2−/− mice (BALB/c-129SJ) generated previously by homologous recombination and backcrossed onto the BALB/c background. Ptpn2ex2−/ex2− mice exhibited growth retardation and a median survival of 32 days, as compared to 21 days for Ptpn2−/− (BALB/c) mice, but the overt signs of morbidity (hunched posture, piloerection, decreased mobility and diarrhoea) evident in Ptpn2−/− (BALB/c) mice were not detected in Ptpn2ex2−/ex2− mice. At 14 days of age, bone development was delayed in Ptpn2−/− (BALB/c) mice. This was associated with increased trabecular bone mass and decreased bone remodeling, a phenotype that was not evident in Ptpn2ex2−/ex2− mice. Ptpn2ex2−/ex2− mice had defects in erythropoiesis and B cell development as evident in Ptpn2−/− (BALB/c) mice, but not splenomegaly and did not exhibit an accumulation of myeloid cells in the spleen as seen in Ptpn2−/− (BALB/c) mice. Moreover, thymic atrophy, another feature of Ptpn2−/− (BALB/c) mice, was delayed in Ptpn2ex2−/ex2− mice and preceded by an increase in thymocyte positive selection and a concomitant increase in lymph node T cells. Backcrossing Ptpn2−/− (BALB/c) mice onto the C57BL/6 background largely recapitulated the phenotype of Ptpn2ex2−/ex2− mice. Taken together these results reaffirm TCPTP's important role in lymphocyte development and indicate that the effects on morbidity, mortality, bone development and the myeloid compartment are strain-dependent.


Introduction
T cell protein tyrosine phosphatase (TCPTP) encoded by PTPN2 is an intracellular tyrosine-specific protein tyrosine phosphatase (PTP) [1]. TCPTP is expressed ubiquitously, but it is most abundant in cells of hematopoietic origin where it serves as a key regulator of hematopoiesis and immune and inflammatory responses [1]. Genome-wide association studies have linked PTPN2 single nucleotide polymorphisms (SNPs) with the development of autoimmune diseases including inflammatory bowel disease, type 1 diabetes and rheumatoid arthritis [2,3,4]. Recent studies have reported that type 1 diabetes-associated PTPN2 intronic SNPs may result in decreased PTPN2 messenger RNA [5]. Moreover, PTPN2 is deleted in 6% [6] of all T cell acute lymphoblastic leukemias, in particular in those overexpressing the TLX1 transcription factor oncogene, promoting proliferation and enhancing cytokine sensitivity [7]. In addition, alterations in TCPTP expression may be associated with several other malignancies and inflammatory diseases [8,9,10,11].
TCPTP's role in hematopoiesis, immunity and inflammation is highlighted by the phenotype of Ptpn2 2/2 mice on a mixed BALB/c-129SJ background [25]; mice appear physically normal albeit slightly smaller at 10-14 days of age, but thereon exhibit growth retardation, a hunched posture, piloerection, decreased mobility and diarrhoea and overt signs of wasting disease, succumbing by 3-5 weeks of age [25]. Ptpn2 2/2 (BALB/c-129SJ) mice are characterized by widespread mononuclear infiltrates in non-lymphoid tissues and elevated serum interferon (IFN) c levels by 19 days of age [26]. Consistent with the inflammatory phenotype, Ptpn2 2/2 (BALB/c-129SJ) mice are more susceptible to endotoxic shock and macrophages are hypersensitive to lipopolysaccharide [26] and IFNc [19], whereas Ptpn2 +/2 (BALB/c-129SJ) mice are hypersensitive to dextran sulphate sodium (DSS) and the development of colitis [27]. Ptpn2 2/2 (BALB/c-129SJ) mice exhibit defects in hematopoiesis, in particular in B cell development and erythropoiesis and develop severe anemia [25]. The defect in B cell development is thought to be attributable, at least in part, to a bone marrow stromal defect and the abnormal secretion of IFNc by bone marrow stromal cells [28]. Ptpn2 2/2 (BALB/c-129SJ) mice develop thymic atrophy associated with a drop in double positive (CD4+CD8+) thymocytes, lymphadenopathy and splenomegaly, the latter being associated with an accumulation of myeloid cells and the sequestration of red blood cells [25].
The Ptpn2 2/2 (BALB/c-129SJ) mice were generated more than ten years ago using a homologous recombination approach that resulted in the deletion of approximately 9 kb of genomic sequence, including 1.5 exons (encoding residues 64-121) from the TCPTP catalytic domain, and inserted a Neomycin (Neo) resistance cassette under the control of a Herpes simplex thymidine kinase promoter for the purposes of ES cell selection [25]. It is now widely appreciated that the insertion of Neo genes can have unintended consequences and reduce the expression of nearby flanking genes and complicate or even confound our understanding of gene function [29,30]. With the advent of Cre/ LoxP recombination, constructs can now be designed for the elimination of Neo resistance genes. Given the importance of PTPN2 in human disease and the complex phenotype in Ptpn2 2/2 (BALB/c-129SJ) mice, we used Cre/LoxP gene targeting to generate Ptpn2 (C57BL/6) null mice without the Neo cassette and compared the phenotype of these mice to that of Ptpn2 2/2 (BALB/c-129SJ) mice backcrossed onto the BALB/c versus C57BL/6 backgrounds. Our results confirm and further characterise the previously reported phenotype and identify straindependent differences in bone development, thymocyte development and myeloid hyperplasia and overall differences in morbidity and mortality.

Targeted disruption of Ptpn2
The Ptpn2 gene is located on chromosome 18 and has 7 exons spread over approximately 60 kb of genomic DNA. To generate a Ptpn2 null mouse using Cre/LoxP targeting, a construct was designed with a PGK (Phosphoglycerate kinase)-Neo cassette flanked with FRT recombination sites inserted downstream of exon 2, and both exon 2 and PGK-neo flanked with LoxP recombination sites for deletion using Cre recombinase; excision of exon 2 would splice exon 1b to exon 3 to induce a frame-shift that would result in early termination (TAA) and a putative 87 residue protein (10,320 Da) incorporating 53 residues from the TCPTP N-terminus (Fig. 1a). The linearised construct was electroporated into C57BL/6 derived Bruce 4 embryonic stem (ES) cells. Correctly targeted Neo resistant clones were injected into blastocysts and high percentage chimeras crossed to C57BL/6J Oz-Cre deleter mice (Ozgene) to excise exon 2 and the Neo cassette ( Fig. 1b) and thereon bred with C57BL/6J mice to eliminate the Cre recombinase (Fig. 1c). The resulting Ptpn2 exon2-deleted mice (Ptpn2 ex2+/ex22 ) were maintained by breeding heterozygotes (Fig. 1d). Ptpn2 ex22/ex22 were born at the expected Mendelian frequency (Table S1). Western blot analyses using antibodies (6F3) to the TCPTP non-catalytic C-terminus confirmed that full length TCPTP protein was not detectable in any of the tissues examined in homozygous Ptpn2 ex22/ex22 mice (Fig. 1e). The hypothetical 87 residue peptide resulting from the splicing of exon 1b to 3 could not be detected with antibodies (6F7; raised against the first 38 residues of TCPTP) to the TCPTP N-terminus (Fig. 1f), consistent with the corresponding mRNA and/or protein being unstable and the mice being null for TCPTP.

Bone development in Ptpn2-deficient mice
Previous studies have indicated that a bone marrow stromal cell deficiency underlies the inflammatory phenotype and overt Figure 1. Generation of Ptpn2 ex2+/ex22 mice. (a) Ptpn2 genomic locus and targeting design. (b-c) Southern blot analysis of Ptpn2 ex2+/ex2+ (+/+) and Ptpn2 ex2+/ex22 (+/2) mice. (b) Exon (Ex) 2 floxed mice mated with Oz-Cre mice for the germline deletion of exon 2 and removal of the Neomycin (Neo) resistance cassette. (c) Ex2-deleted mice were subsequently mated with C57BL/6 mice for the elimination of the Cre transgene. (d) PCR analysis of Ptpn2 ex2+/ex2+ (+/+), Ptpn2 ex2+/ex22 (+/2) and Ptpn2 ex22/ex22 (2/2) mice. (e-f) Immunoblot analysis of TCPTP expression in lymphoid and nonlymphoid tissues detected with antibodies to the TCPTP non-catalytic C-terminus (6F3) and TCPTP N-terminus (6F7; raised against the first 38 residues of TCPTP); a non-specific (n.s.) protein detected by 6F7 is indicated. Results are representative of three independent experiments. doi:10.1371/journal.pone.0036703.g001 morbidity in Ptpn2 2/2 (BALB/c-129SJ) mice [25]. More recent studies have shown that TCPTP deficiency in osteoblasts (C57BL/ 6J-129SJ) increases osteoclast activity in vitro, without significantly affecting the osteoblast or osteoclast number or bone formation/ volume in vivo [31]. We characterized the bones of Ptpn2 2/ 2 (BALB/c) and Ptpn2 ex22/ex22 mice to determine whether the differences in morbidity might be associated with differences in bone development. Hind legs were isolated from 14 day-old Ptpn2 2/2 (BALB/c) and corresponding Ptpn2 +/2 and Ptpn2 +/+ littermates prior to the onset of overt morbidity [25,26], as well as 14 day-old Ptpn2 ex22/ex22 and Ptpn2 ex2+/ex2+ littermate mice. Legs were fixed in 3.7% formaldehyde and processed for bone histology and histomorphometric analysis as described previously [32]. 14 day-old Ptpn2 2/2 (BALB/c) mice had significantly smaller skeletons, reflected in reduced femoral length and width, and delayed development of the secondary ossification centre, indicated by the high level of cartilage remaining in this zone (Fig 3). In addition, in the proximal tibial metaphysis, a region of bone remodeling, Ptpn2 2/2 (BALB/c) mice had increased trabecular bone volume, characterized by high trabecular number (TbN) and trabecular thickness (TbTh) and reduced trabecular separation (TbSp) ( Fig. 3; Table 1). A lower level of remodeling was indicated by reduced osteoid deposition, and a significant reduction in osteoblast surface, along with a slight, but not statistically significant reduction in osteoclast surface ( Table 1). Interestingly in 21 day-old Ptpn2 2/2 (BALB/c) mice, the delay in the secondary ossification centre was no longer evident and the increase in trabecular bone was less obvious (Fig. 3). In contrast to Ptpn2 2/2 (BALB/c), we found slightly, but not significantly lower femoral length in 14 day-old Ptpn2 ex22/ex22 mice, but no overt difference in femoral width or trabecular bone volume ( Fig. 3; Table 2). Histological assessment indicated that Ptpn2 2/2 (C57BL/6) mice also did not exhibit any overt difference in size, nor increased trabecular bone at 14 days of age ( Fig. 3). At 21 days of age Ptpn2 ex22/ex22 bones were overtly smaller, and this was associated with a striking reduction in the growth plate width; again there was no increase in trabecular bone (Fig. 3). Taken together, these results point towards TCPTP-deficiency resulting in strain-dependent differences in bone development and remodeling.

Body composition in Ptpn2-deficient mice
Given the growth retardation, but discordant morbidity in Ptpn2 2/2 (BALB/c) versus Ptpn2 2/2 (C57BL/6) and Ptpn2 ex22/ ex22 mice, we assessed the impact of TCPTP deficiency on body composition by dual energy X-ray absorptiometry (DEXA) at 18 and 28 days as appropriate (Tables 3-4). Lean mass, fat mass and bone mineral content and density were assessed in Ptpn2 ex22/ex22 and Ptpn2 2/2 (C57BL/6) mice and compared to that of Ptpn2 2/ 2 (BALB/c) mice and their corresponding heterozygous and/or wild type littermates. Lean mass, fat mass and bone mineral content but not bone mineral density, were reduced in Ptpn2 ex22/ ex22 , Ptpn2 2/2 (C57BL/6) and Ptpn2 2/2 (BALB/c) mice when compared to their corresponding controls, consistent with the    overt growth retardation ( Tables 3-4). However, lean masses normalized to total body weight were unaltered in Ptpn2 ex22/ex22 , Ptpn2 2/2 (C57BL/6) and Ptpn2 2/2 (BALB/c) mice when compared to controls, in line with the mice being proportionately smaller and the difference in growth being developmental in nature. Relative fat masses were decreased in Ptpn2 ex22/ex22 , Ptpn2 2/2 (C57BL/6) and Ptpn2 2/2 (BALB/c) mice when compared to controls, consistent with TCPTP deficiency affecting energy balance, whereas bone mineral densities normalized to body weight were increased (Tables 3-4); at 18 days of age the increase in bone mineral density was greatest in Ptpn2 2/2 (BALB/ c) mice consistent with the strain-dependent differences in bone development as revealed by histology and histomorphometry ( Fig. 3; Table 1). Overall there were no overt differences in body composition between Ptpn2 ex22/ex22 , Ptpn2 2/2 (C57BL/6) and Ptpn2 2/2 (BALB/c) mice, consistent with the strain-dependent differences in morbidity and mortality being independent of gross body composition.
These results indicate that the effects of TCPTP deficiency on splenomegaly, myeloid development and lymphodenopathy are strain-dependent.

B cell development in Ptpn2-deficient mice
We next assessed B cell development in 18 day-old Ptpn2 2/2 (BALB/c), 28 day-old Ptpn2 2/2 (C57BL/6) and 29 day-old Ptpn2 ex22/ex22 mice (versus corresponding littermates). B cells originate from precursors in the bone marrow and undergo sequential maturation and V(D)J recombination developing into immature B cells that express B cell receptor and membranebound IgM [35]. Immature B cells emigrate to the spleen where they mature further into three broad populations according to anatomical localization, phenotypic characterization and function: follicular B cells (IgM lo IgD hi B220 hi ), marginal zone B cells (IgM hi IgD lo B220 hi ) and B-1 cells (IgM hi IgD lo B220 lo ) [35]. Previous studies have shown that B cell intrinsic defects and bone marrow stromal abnormalities in TCPTP-deficient mice lead to an early block in B cell development [25,28]. Consistent with this we found a dramatic reduction in total B220 + B cells, IgM 2 IgD 2 -B220 lo progenitor B cells, IgM hi IgD lo B220 lo immature B cells and IgM int IgD hi B220 hi mature recirculating B cells in bone marrow and decreased splenic marginal zone B cells, follicular B cells and B-1 cells (Fig. 8a-b). Although bone marrow B cell development was perturbed in all three mouse lines, the impact of TCPTP deficiency on the development of splenic B cells was most dramatic in Ptpn2 2/2 (C57BL/6) and Ptpn2 ex22/ex22 mice (Fig. 8b).  Conversely, LN B cell subsets were most dramatically altered in Ptpn2 2/2 (BALB/c), with no overt differences seen in Ptpn2 ex22/ ex22 mice (Fig. 8c).

Erythropoiesis in Ptpn2-deficient mice
Ptpn2 2/2 (BALB/c-129SJ) mice develop severe anemia by 21 days of age associated with defective bone marrow erythropoiesis [25]. Erythropoiesis is a dynamic process that involves multiple differentiation steps and progression from early progenitor cells to enucleated red blood cells. Pluripotent hematopoietic stem cells and multipotent progenitor cells in the bone marrow give rise to committed erythroid precursors, which sequentially develop into mature erythrocytes [36]. We found that CD117 + Ter119 2 progenitor cells and CD117 2 Ter119 + erythroid cells (early proerythroblast to mature erythrocyte) were decreased by more than 80% in Ptpn2 2/2 (BALB/c), Ptpn2 2/2 (C57BL/6) and Ptpn2 ex22/ ex22 mice when compared to their corresponding littermates (Fig. 9a). Circulating erythrocytes were also reduced, but this was most dramatic in Ptpn2 2/2 (BALB/c) mice (Fig. 9b). Interestingly, the decrease in erythroid cells was greater than the decrease in erythrocytes (Fig. 9), consistent with tissues such as the spleen and liver compensating for the defective bone marrow erythropoiesis. These results indicate that erythroid development is defective in TCPTP-deficient mice.

Discussion
In this study we have generated mice that are globally deficient for the tyrosine phosphatase TCPTP using Cre/LoxP recombination to remove both Ptpn2 exon 2 and the Neomycin resistance cassette used for ES cell selection. The phenotype of these mice reaffirms that reported previously for mice generated by homologous recombination, where the Neo cassette was left in place [25]. Thus, our studies indicate that the morbidity and mortality and the perturbations in hematopoiesis and erythropoiesis in Ptpn2 2/2 (BALB/c) can be ascribed to a deficiency in TCPTP, rather than extrinsic influences that may arise from the presence of the Neo resistance gene. Furthermore, our studies define strain-dependent differences in morbidity and mortality and differences in thymocyte, myeloid and bone development.
Global TCPTP-deficiency was associated with differences in morbidity and mortality on the BALB/c versus C57BL/6 background strains. The outward signs of morbidity in Ptpn2 2/2 (BALB/c) mice, including hunched posture, piloerection, decreased mobility, eyelid closure and diarrhoea were not evident in the corresponding mice on the C57BL/6 background. Moreover, lifespan was prolonged in TCPTP-deficient on the C57BL/6 background. The underlying reason(s) for the strain-dependent differences in morbidity and mortality is/are not clear. One TCR-Vb3, -5, -6 and -8 T cells determined. (c) Thymocytes from the indicated 18 day-old mice were stained with fluorochrome-conjugated antibodies against CD4, CD8, TCRb and CD69 and analyzed by flow cytometry. Cells were gated for the different developmental stages according to the expression of the positive selection markers TCRb and CD69. Results in (a-c) are means 6 SEM for the indicated number of mice and are representative of at least two independent experiments; significance determined using a two-tailed Mann-Whitney U Test; *p,0.05 ** p,0.01 *** p,0.001. doi:10.1371/journal.pone.0036703.g004 possible cause may be differences in the myeloid compartment. Although growth retardation and defects in lymphocyte and erythroid development were evident irrespective of the background strain, the impact on the myeloid compartment was distinct. Despite bone marrow cellularity being decreased in both 28 day-old C57BL/6 and 18 day-old BALB/c strain TCPTPdeficient mice (in line with their smaller size), only C57BL/6 TCPTP-deficient mice exhibited a corresponding decrease in bone marrow CD11b + cells. Thus, monocytic/granulocytic development may be increased in Ptpn2 2/2 (BALB/c) mice. This was accompanied by a dramatic increase in CD11b + Gr1 + , but not CD11b + Gr1 2 cells in the spleens of Ptpn2 2/2 (BALB/c) mice. The anti-granulocyte receptor 1 (Gr1) antibody RB6-8C5 binds to Ly6G on the surface of granulocytes and Ly6C present on both granulocytes and monocytes/macrophages and some lymphocytes. Two major monocytic subsets exist in the periphery, the 'inflammatory' and 'circulating' monocytes [37]. Inflammatory monocytes express Ly6C (Gr1 + ) and migrate to the spleen and inflamed tissues, where they become macrophages, whereas circulating monocytes are Gr1 2 and serve to replenish resident tissue macrophages. The dramatic increase in granulocytes and/or inflammatory macrophages in the spleens of Ptpn2 2/2 (BALB/c) mice is consistent with inflammation and previous studies establishing the development of progressive systemic inflammatory disease in Ptpn2 2/2 (BALB/c-129SJ) mice [25].
Our studies also define strain-dependent differences in thymic atrophy and thymocyte development that might contribute to the differences in morbidity in BALB/c versus C57BL/6 strain TCPTP-deficient mice. In particular, thymic atrophy and the overt decrease in thymic cellularity were evident sooner in the Ptpn2 2/2 (BALB/c) mice. The reasons for the strain-dependent temporal differences in thymic atrophy are not clear. The decrease in thymocyte cellularity in BALB/c and C57BL/6 strain TCPTPdeficient mice might be attributable at least in part to the nonspecific depletion of DP thymocytes by inflammatory cytokines such as TNF and IFNc that are increased in Ptpn2 2/2 (BALB/c-129SJ) mice [26] and have been linked previously with the induction of DP apoptosis in vivo [38]. However, this might also be a consequence of defective lymphoid progenitor development in the bone marrow. In future studies it will be important to determine whether there are strain-dependent differences in inflammation and lymphoid progenitor development in Ptpn2 2/ 2 mice. Nevertheless, despite the decreased cellularity, we found that TCPTP-deficiency enhanced both positive and negative selection and overall resulted in increased SP/DP ratios. The enhanced positive selection was evident in particular in young Ptpn2 ex22/ex22 and reflected by an increase in peripheral T cell numbers. These results are consistent with our recent analyses of thymocyte/T cell-specific TCPTP-deficient (Lck-Cre;Ptpn2 lox/lox ) mice, where positive selection and SP/DP ratios are also increased [39]. Although T cell-specific TCPTP deficiency results in inflammation and autoimmunity in aged mice [24], it is possible that the enhanced T cell numbers in young Ptpn2 ex22/ex22 might be beneficial.
Phenotypic differences in different strains of mice carrying targeted null mutations have been reported previously and linked with the existence of second-site modifier genes that affect penetrance [40,41,42,43,44]. In particular, mice lacking the retinoblastoma (Rb)-related p107 gene develop growth deficiency on the BALB/c, but not C57BL/6 background strains [40]. Indeed, the overall phenotype of p107 2/2 (BALB/c) mice is highly reminiscent of that seen in Ptpn2 2/2 (BALB/c) mice. In addition to postnatal growth retardation, p107 2/2 (BALB/c) mice have a higher incidence of morbidity and exhibit myeloid hyperplasia [40] and a suppression of B-lymphopoiesis and erythropoiesis [45]. The myeloid hyperplasia in p130-deficient mice is linked with perturbations in the bone marrow microenvironment [45]. Ptpn2 2/2 (BALB/c-129SJ) mice also exhibit bone marrow stromal cell defects associated with increased INFc production [26,28]. The defect(s) in the bone marrow microenvironment in Ptpn2 2/2 Figure 6. Thymoycte development in 28 day-old Ptpn2 ex22/ex22 and Ptpn2 2/2 (C57BL/6) mice. Thymi from the indicated littermate mice were weighed using an analytical balance and normalised to body weight. Thymocytes were stained with fluorochrome-conjugated antibodies against CD4 and CD8 and analyzed by flow cytometry. Cells were gated for the DP, SP and DN stages and absolute numbers and the indicated ratios determined. Results shown are means 6 SEM for the indicated number of mice and are representative of at least two independent experiments; significance determined using a two-tailed Mann-Whitney U Test; *p,0.05 *** p,0.001. doi:10.1371/journal.pone.0036703.g006 (BALB/c-129SJ) mice contribute(s) to morbidity and mortality and has/have been shown to be responsible for the block in B cell development [25,28]. Although we did not assess if TCPTPdeficiency results in strain-dependent differences in the bone marrow stroma, we did observe a striking difference in bone development in Ptpn2 2/2 (BALB/c) versus Ptpn2 ex22/ex22 mice.
At 14 days of age Ptpn2 2/2 (BALB/c) mice had smaller skeletons due to a delay in bone development, as reflected by the large volume of unresorbed cartilage in the secondary ossification centre (at the epiphysis). Moreover, we found that trabecular bone was increased significantly in Ptpn2 2/2 (BALB/c) mice, comprising mainly of woven bone rather than more mature lamellar bone, consistent with the developmental delay. Our findings are consistent with global TCPTP deficiency delaying bone development and decreasing bone turnover on the BALB/c background. Interestingly, in 21 day-old Ptpn2 2/2 (BALB/c) mice the delay in the secondary ossification was no longer evident and only modest effects on trabecular bone were seen. Therefore, these results are not inconsistent with a recent study that has reported that 21 day-old Ptpn2 2/2 (BALB/c) mice exhibit spontaneous synovitis associated with synovial inflammation and increased osteoclast density and decreased rather than increased bone volume [46]. The overt differences we observed in bone development in Ptpn2 2/2 (BALB/c) mice were not present in Ptpn2 ex22/ex22 mice at 14 days of age. Moreover in contrast to 21 day-old Ptpn2 2/2 (BALB/c) mice, TCPTP deficiency on the Figure 7. Splenomegaly, myeloid development and lymphadenopathy in Ptpn2-deficient mice. (a) Spleen weights from 18 day-old Ptpn2 2/2 (BALB/c) and 28 day-old Ptpn2 ex22/ex22 and Ptpn2 2/2 (C57BL/6) mice and corresponding wild type littermate mice. (b) Splenocytes or (c) bone marrow (tibia and femur) cells were stained with fluorochrome-conjugated antibodies to CD11b and Gr1 and analysed by flow cytometry; absolute numbers of CD11b + , CD11b + Gr1 + and CD11b + Gr1 2 were determined (d) Weights of pooled peripheral lymph nodes from 18 day-old Ptpn2 2/2 (BALB/c) and 28 day-old Ptpn2 ex22/ex22 and Ptpn2 2/2 (C57BL/6) mice and corresponding wild type littermate mice. Results shown in (a-d) are means 6 SEM for the indicated number of mice and are representative of at least two independent experiments; significance determined using a two-tailed Mann-Whitney U Test; *p,0.05 ** p,0.01 *** p,0.001. doi:10.1371/journal.pone.0036703.g007 C57BL/6 background resulted in a striking narrowing of the growth plate at 21 days of age; an influence that would cause a slowing of longitudinal growth and might therefore be responsible for the smaller skeletons in these mice. Therefore, although TCPTP-deficiency results in smaller skeletons irrespective of strain by 21 days of age, the underlying causes might be different and strain-dependent. A similar growth plate phenotype to that of the Ptpn2 ex22/ex22 mice has been noted previously in growth hormone (GH) receptor deficient mice [47]. Interestingly, mice that lack TCPTP specifically in neuronal cells (Nes-Cre;Ptpn2 lox/lox ) are also runted and this is associated with decreased circulating GH, which, amongst other influences, affects post-natal growth [39].
It is well established that the bone microenvironment plays a central role in hematopoietic stem cell fate [48,49]. Further studies should focus on delineating the precise cellular and molecular mechanisms that underlie the strain-dependent differences in bone development and potentially bone marrow stroma in Ptpn2 2/2 mice and the relative contributions of bone to the perturbations in hematopoiesis and erythropoiesis and the morbidity and mortality that are associated with global TCPTPdeficiency.

Ethics statement
All experiments were performed in accordance with the NHMRC Australian Code of Practice for the Care and Use of Animals and approved by the Monash University School of Biomedical Sciences Animal Ethics Committee (Ethics number SOBSB/B/2008/53). Figure 8. B cell development in Ptpn2-deficient mice. Bone marrow cells (pooled from one tibia and one femur), splenocytes or pooled peripheral lymph node cells from 18 day-old Ptpn2 2/2 (BALB/c), 28 day-old Ptpn2 ex22/ex22 and 29 day-old Ptpn2 2/2 (C57BL/6) mice and their corresponding littermates (a-c) were stained with fluorochrome-conjugated antibodies to CD45R(B220), IgM and IgD and analysed by flow cytometry. Absolute numbers of progenitor (B220 lo IgM lo IgD lo ), immature (B220 lo IgM hi IgD lo ), mature (B220 hi IgM int IgD hi ) B cells and B1 (B220 lo IgM hi IgD lo ), follicular (B220 hi IgM lo IgD hi ) and marginal zone (B220 hi IgM hi IgD lo ) B cells were determined. Results shown are means 6 SEM for the indicated number; significance determined using a two-tailed Mann-Whitney U Test; *p,0.05 ** p,0.01 *** p,0.001.

Mice
We maintained mice on a 12 h light-dark cycle in a temperature-controlled high barrier facility (Monash University ARL) with free access to food and water. Ptpn2 2/2 mice on a 129sv6BALB/c mixed background have been described previously [25] and were backcrossed onto a BALB/c background for eight generations. Ptpn2 2/2 (BALB/c) were backcrossed onto a C57BL/6 background for seven generations. Ptpn2 2/2 (BALB/c) and Ptpn2 2/2 (C57BL/6) mice were genotyped as described previously [25]. Littermates and aged-matched mice from heterozygous breeding pairs were used in all experiments.
Generation and genotyping of Ptpn2 ex22/ex22 Mice Ptpn2 ex2+/ex22 mice were generated by Ozgene (Perth, Australia). The targeting construct was produced using C57BL/6 genomic DNA and the Ozgene PelleNeo vector and incorporated LoxP sites flanking both exon (Ex) 2 and a PGK-Neo-selection cassette flanked by FRT sites inserted downstream of exon 2 (Fig. 1). The linearised construct was electroporated into Bruce 4 (C57BL/6J) ES cells and correctly targeted G418 resistant clones identified by Southern blotting and injected into blastocysts for the generation of male chimeras that were mated with C57BL/6J mice to produce Ptpn2 ex2+/ex22 offspring. Deletion of exon 2 and removal of the Neo cassette were achieved by crossing with Oz-Cre deleter mice (Ozgene) and the resulting progeny bred with C57BL/6 mice for the elimination of the Cre transgene; deletion of exon 2, removal of Neo and the elimination of Cre were monitored by Southern blotting. Ptpn2 ex2+/ex22 mice were genotyped by PCR: forward primer P1 59TGCAGT-TATGGTTTTCCTCAGTCCC39 and reverse primer P2: 59ACAGT GCTGGTTGCTGTTTAGCCTC39.

Flow cytometry
Freshly isolated thymi, spleens and peripheral lymph nodes were homogenized by gently compressing them between two frosted glass slides or with the sterile end of a plunger and washed with ice-cold PBS supplemented with 2% (v/v) fetal bovine serum (FBS; CSL, Australia); cell suspensions were resuspended further with an 18G needle. Bone marrow was isolated by flushing tibias and femurs with ice cold PBS supplemented with 0.2% BSA with a 26G needle. The resulting marrow was then gently resuspended using a 22G needle. Cell suspensions were recovered by centrifugation (3006 g, 5 min at 4uC), red blood cells removed Figure 9. Erythrocyte development in Ptpn2-deficient mice. (a) Bone marrow cells (pooled from one tibia and one femur) from 18 day-old Ptpn2 2/2 (BALB/c) and 28 day-old Ptpn2 ex22/ex22 and Ptpn2 2/2 (C57BL/6) mice and corresponding littermate wild type control mice were stained with fluorochrome-conjugated antibodies to CD117 and Ter119 and analysed by flow cytometry. Absolute numbers of progenitor (CD117 + Ter119 2 ) and erythroid (CD117 2 Ter119 + ) cells were determined. (b) Red blood cells were quantified (identified in the forward and side scatter according to size and granularity) by flow cytometry. Results shown are means 6 SEM for the indicated number of mice and are representative of two independent experiments; significance determined using a two-tailed Mann-Whitney U Test; *p,0.05 ** p,0.01 *** p,0.001. doi:10.1371/journal.pone.0036703.g009 using a red blood cell lysing buffer (Sigma-Aldrich, St Louis, MO) and cell counts (4-15 mm) determined with a Z2 Coulter Counter (Beckman Coulter, Fullerton, CA). For cell surface staining, 3610 6 lymphocytes were resuspended in PBS/2% FBS containing the fluorochrome-conjugated antibodies and incubated on ice for 20 min. Cells were washed once in PBS/2% FBS to remove unbound antibodies and analyzed on a LSRII flow cytometer (BD Biosciences, San Jose, CA). For the quantification of red blood cells, mice were bled and a known number of Calibrite TM Beads (BD Biosciences, San Jose, CA) added to a defined volume of diluted blood and analysed by flow cytometry (LSRII). Red blood cells were identified in the forward and side scatter according to their size. Data was analyzed using FACSDiVa (BD Biosciences, San Jose, CA) or FlowJo7 (Tree Star Inc., Ashland, OR) software.

Histomorphometry
Tibial specimens were fixed in 4% paraformaldehyde/PBS. For histomorphometry, tibiae were embedded in methylmethacrylate [32]. Five mm sagittal sections were stained with toluidine blue, von Kossa stain, and safranin O and histomorphometry was carried out in the secondary spongiosa of the proximal tibia as previously described (OsteoMeasure TM , OsteoMetrics, Decatur, GA) [32].

Body composition
Body composition was measured by DEXA (Lunar PIXImus2; GE Healthcare) and analysed using PIXImus2 software; the head region was excluded from analyses.

Statistical Analyses
Statistical analyzes were performed using the nonparametric, unpaired Mann-Whitney U test and the Long Rank (Mantel-Cox) test using Graphpad Prism (San Diego, CA) software. P values of ,0.05 were considered significant.