Inbred C57BL/6 mice and B6.Cg-Tg(CD4-cre)1Cwi mice were purchased from Taconic Farms (Hudson, NY, USA). Inbred C57BL/6 mice, B6.SJL-Ptprca Pepcb/BoyJ, B6.Cg-Tg(Lck-cre)3779Nik/J, B6.129S7-Rag1^Tm1Mom/J and B6.129P2-Ptpn6^tm1Rsky/J (SHP-1^fl/fl) were purchased from The Jackson Laboratory. Ptpn11^fl/fl (SHP-2^fl/fl) mice with LoxP recombination sites flanking Ptpn11 were previously described (18) and bred with CD4-Cre mice and Lck-Cre mice. SHP-2^fl/fl-CD4-Cre mice were crossed to R26R-EFYP mice to produce SHP-2^fl/fl-CD4-Cre-ROSA^EYFP mice. Jα18^−/− mice were bred, crossed to B6 (>12 generations). SHP-1^fl/fl mice were bred with CD4-Cre recombinase mice. All mice were bred in pathogen-free breeding facilities at Brown University. All experiments were conducted in accordance with institutional guidelines for animal care.

Mice were sacrificed by isofluorane treatment. Cardiac puncture was performed prior to the harvesting of the organs. Liver was perfused before harvesting. Spleens were dissociated and lymphocytes were enriched using Lympholyte Cell Separation Media (Cederlane). Livers were dissociated using a gentleMACS Dissociator (Miltenyi Biotec) and lymphocytes were enriched using a 40–70% discontinuous Percoll gradient (GE Healthcare) as previously described (19). Thymi were dissociated and thymocytes were washed prior to analysis.

Cells were stained with fluorochrome-conjugated monoclonal antibodies and 2.4G2 blocking antibody and incubated in the dark for 30 min on ice. For CD1d tetramer staining, cells were incubated in the dark for 15 min at room temperature followed by 15 min on ice. Flow cytometric analysis was performed on a FACSAria (BD Biosciences) or a MACSQuant (Miltenyi Biotec). Data were analyzed using FlowJo (Tree Star).

TCRβ-FITC, HSA-FITC, CD4-FITC, TCRβ-PE, Ly49C/I/F/H-PE, TCRβ-PerCPCy5.5, Ly49G2-PerCPeFluor710, CD62L-PE-Cy6, NK1.1-PE-Cy7, KLRG1-APC, CD25-APC, CD19-PE, CD44-APCeFluor780, CD8α-eFluor450, and CD3-eFluor450 were purchased from ebioscience. Ly49G2-FITC, Ly49C/I-FITC, Ly49A/D-PE, and CD4-PerCP were purchased from BD Pharmingen. BV421-CD127, BV510-TCRβ, BV570-CD45, BV605-CD3, BV605-CCR6, and BV785-NK1.1 were purchased from Biolegend. PE- and APC-conjugated α-GalCer loaded CD1d tetramer was prepared in our laboratory. CD1d, M45 and M57 tetramers were kindly provided by the National Institute of Allergy and Infectious Disease MHC Tetramer Core Facility at Emory University (Atlanta, GA, USA).

Splenocytes from SHP-2 heterozygous (fl/+) and knockout (fl/fl) animals were sorted using a FACSAria to greater than 95% efficiency. TCRβ⁺ and TCRβ⁻ populations were collected and lysed in Lysis buffer. Cellular lysates were added to SDS/sample buffer, boiled, and run on a Ready Gel 4–20% Tris-HCl (Bio-Rad, Hercules, CA, USA) with 1× SDS-PAGE running buffer. The gel was blotted against a nitrocellulose membrane (Bio-Rad) using 1× transfer buffer with 10% methanol. After blocking in a 5% milk solution for 1 h at room temperature, rabbit α-mouse SHP-2 (Abcam) and β-actin (Ambion) antibodies were incubated with the membrane overnight at 4°C. HRP-conjugated-donkey α-rabbit IgG secondary antibody (Jackson Immunoresearch, Westgrove, PA, USA) was incubated with the membrane for 1 h at room temperature. Membranes were developed with Super signal Chemiluminescent Substrate (Pierce) and photographed using a ChemiDoc XRS + System (Bio-Rad).

X-ray images of the entire skeleton were obtained immediately after euthanasia using a Faxitron X-ray system (Wheeling). After fixation in 4% paraformaldehyde, μ-CT images of skeletal tissues were scanned with a desktop microcomputer graphic imaging system (μ-CT40, Scanco Medical AG). For μ-CT, 3D image acquisitions took 36 min with a beam parameter of 70 kVp.

Mice were euthanized at the indicated ages, and femurs, tibiae, and paws were removed and fixed in 4% paraformaldehyde overnight at 4°C. Postnatal skeletal tissues were decalcified in 0.5 M EDTA before embedding. Tissue sections (5 µm) were stained with hematoxylin and eosin. Some sections were separately stained with Safranin O and Fast Green.

American Type Culture Collection Strain #VR1399 and RVG-102 clone of MCMV recombinant for GFP under the promoter of the immediate early gene-1 (i.e., 1) (a gift of Dr. Hamilton, Duke University) were prepared from salivary glands as previously described (20, 21). On day 0, mice were infected intraperitoneally with 5 × 10⁴ plaque-forming units (pfu) of MCMV. Mice were sacrificed for analysis at 7 days and 8–12 weeks postinfection.

A method described by Stankovic et al was used (22). Briefly, CD1d tetramer-positive cells were sorted using a FACSAria to greater than 95% efficiency. A 96-well plate was coated with 0.5 μg/well purified α-mouse CD3ε and CD28 antibodies (eBioscience) and iNKT cells were plated at a concentration of 10,000 cells per well. Supernatant was collected 24 h poststimulation. Cytokine production was quantified by Cytometric Bead Array (BD Biosciences). Samples were run on a FACSAria and data were analyzed using FCAP Array Software (BD Biosciences).

Recipient mice were non-lethally irradiated with 7.5 Gy and placed on oral Sulfamethoxazole and Trimethoprim (Hi-tech Pharmacal). On day 3 postirradiation, donor mice were sacrificed, thymi were harvested, and thymocytes were extracted and counted. CD8⁺ T cells were depleted using anti-CD8 magnetic beads via an AutoMACS. Remaining cells were stained for 10 min at 37°C in the dark with Cell Proliferation Dye eFluor450 (eBioscience) in PBS at a labeling concentration of 10 µM. 8–10 million cells were injected into each irradiated recipient. Six days postinjection, mice were sacrificed and the spleen and liver were harvested. Lymphocytes were extracted, stained, run on a MACSquant, and analyzed using FlowJo.

Data sets were analyzed for statistical significance using unpaired, two-tailed, Student’s t-tests. Values of p < 0.05 were considered significant. Graphs and statistics were generated and calculated using Prism (GraphPad).

SHP-2 floxed/floxed (fl/fl) mice (18) were crossed to mice carrying a transgene for Cre recombinase driven by the cd4 enhancer/promoter/silencer (23, 24). As a result, the Ptpn11 gene encoding SHP-2 is deleted at the DN4/DP stages of T cell development when CD4 is first expressed. For simplicity, we will refer to SHP-2^fl/fl-CD4-Cre as SHP-2^−/− and their littermate controls, SHP-2^fl/+-CD4-Cre, as wild type (WT). To verify that SHP-2 was efficiently deleted in T cells, splenic T cells and non-T cell populations from WT and SHP-2^−/− animals were sorted for protein analysis by Western blot using α-mouse SHP-2 primary antibody. Expression of SHP-2 in T cells was strongly reduced to undetectable levels SHP-2^−/− mice (Figure 1A).

We then analyzed thymic T cell development in our conditional knockout mice. We found that WT and SHP-2^−/− mice had comparable frequencies of DP (CD4⁺CD8⁺), single-positive (CD4⁺ or CD8⁺), and DN (CD4⁻CD8⁻) thymic lymphocytes (Figure 1B). Furthermore, no obvious T cell defects were observed in peripheral compartments. The number of splenic T cells (Figure 1C), splenic CD8⁺ T cells (Figures S1A,B in Supplementary Material) and CD4⁺ T cells (data not shown) were comparable between SHP-2^−/− mice and WT controls. However, not all T cells express inhibitory receptors, so the absence of SHP-2 could potentially affect the development of small subsets of T cells known to express these receptors. These include iNKT cells, Ly49⁺CD8⁺ T cells with a memory phenotype (CD44^hiCD122⁺Ly6C⁺) (25), and exhausted T cells during chronic infection (26). To this end, we first examined CD8⁺ T cells with a memory phenotype by staining for CD44 and CD62L. No significant difference in the frequency of effector memory (CD44^hiCD62L^lo) CD8⁺ T cells between WT and SHP-2^−/− mice was observed (Figure S1C in Supplementary Material). SHP-2-deficiency also did not affect the frequency of KLRG1⁺ effector CD8⁺ T cells in all organs we tested (Figure S1D in Supplementary Material). Ly49 antibody cocktail staining revealed that Ly49 (Ly49C/I/F/H) expression did not substantially differ in the spleen, blood, or liver (Figure S1E in Supplementary Material). When analyzed separately, only in the case of the Ly49G2 receptor did we find a modest decrease of hepatic Ly49G2⁺ CD8⁺ T cells in SHP-2^−/− mice (Figure S1F in Supplementary Material).

As previously mentioned, iNKT cells also express a variety of inhibitory receptors such as Ly49 (27). While the role of inhibitory receptors on iNKT cells has not been defined, it has been suggested that they control iNKT cell autoreactivity and tolerance (28, 29). Due to the association of these receptors with SH2 domain-containing phosphatases, we hypothesized that the absence of SHP-2 might affect iNKT cell development. Because iNKT cells are positively selected by CD1d-expressing CD4⁺CD8⁺ cortical thymocytes (30), SHP-2 is absent in both iNKT cell precursors and DP cells mediating positive selection in our conditional knockout mice. However, CD1d tetramer staining showed no significant difference in the absolute number of iNKT cells (Figures S2A,B in Supplementary Material). Additionally, no defects were observed in iNKT cell frequency among each stage of thymic maturation, where the majority of cells progressed to mature Stage III iNKT cells (Figure S2C in Supplementary Material). The TCR Vβ chain repertoire and Ly49 expression profile of iNKT cells from WT and SHP-2^−/− mice were also comparable (data not shown). Collectively, these results indicate that SHP-2 is dispensable during iNKT cell development and maturation.

It has been previously reported that conditional deletion of SHP-2 by placing Cre recombinase under the control of Lck promoter results in defective conventional T cell development (8). As we did not find a T cell defect when SHP-2 was deleted by CD4-Cre, we also crossed SHP-2^fl/fl mice to Lck-Cre transgenic mice, another T cell-specific Cre recombinase mouse line. Similarly to the SHP-2^fl/fl-CD4-Cre mice, no developmental defect was observed in T cells in the thymus (Figure 2A) or in the periphery (Figure 2B). Altogether the data using both SHP-2^fl/fl-CD4-Cre mice and SHP-2^fl/fl-Lck-Cre mice demonstrate that SHP-2 is globally dispensable for T cell development.

We next sought to determine whether deletion of SHP-2 affected iNKT cell and T cells functions. Because activated iNKT cells are known to rapidly secrete large amounts of cytokines upon activation, we quantified cytokine production following in vitro TCR cross-linking as previously described (22). Normalization of three independent experiments revealed that iNKT cell cytokine production was unaffected by SHP-2 absence (Figure S2D in Supplementary Material). More specifically, IL-2, IL-4, IFN-γ, and TNF-α production by SHP-2-deficient iNKT cells did not differ significantly from WT controls. To assess TCR-mediated cytokine production in an in vivo setting, we injected WT and SHP-2^−/− mice intraperitoneally with 2 µg of α-GalCer. α-GalCer is the prototypical ligand for iNKT cells and leads to vigorous cytokine production and proliferation (20). Because iNKT cell TCR cell surface expression is reduced after activation, we analyzed iNKT cells at early time points. At 2 h poststimulation, intracellular cytokine staining showed similar frequencies of IL-4⁻IFN-γ⁻, IL-4⁺, IFN-γ⁺, and IL-4⁺IFN-γ⁺ iNKT cells between WT and SHP-2^−/− mice (Figure S2E in Supplementary Material). To evaluate whether the absence of SHP-2 causes an iNKT cell proliferative defect, we intravenously injected CD8-depleted SHP-2^−/− (CD45.2⁺) and congenic WT (CD45.1⁺) thymocytes into sublethally irradiated Jα18^−/− mice. Proliferation cycles were quantified by Cell Proliferation Dye eFluor450. In these experiments, SHP-2^−/− and congenic WT iNKT cells proliferated to an equal extent in the leukopenic environment (Figure S2F in Supplementary Material). These results suggest that SHP-2 does not affect iNKT cell ability to proliferate.

We next asked whether SHP-2 affected the formation of effector CD8⁺ T cells during infection. Using the MCMV infection model, we examined CD8⁺ T responses at early and late time points (7 days and 8–12 weeks postinfection, respectively). We found the frequencies of CD8⁺ T cells, CD44^hiCD62L^lo effector memory CD8⁺ T cells, and KLRG1⁺CD8⁺ T cells to be nearly identical between WT and SHP-2^−/− mice at both time points (Figure 3). Several MCMV epitopes have been described in C57BL/6 mice including M38, M139, M45, and M57 (21). CD8⁺ T cells specific to M38 and M139 are classified as inflationary cells (31), while CD8⁺ T cells specific for M45 and M57 epitopes have been shown to differentiate into a classical central memory phenotype (31, 32). We simultaneously examined CD8⁺ T cells specific for M45 and M57 epitopes, which expand during acute infection and have been associated with long-term protection against MCMV (21). On day 7 postinfection, no significant difference in the number of effector Ag-specific CD8⁺ T cells was observed (Figure 3B). Furthermore, at 10 weeks postinfection, we found that SHP-2 deficiency did not affect the formation of Ag-specific memory CD8⁺ T cells (Figure 3C,D). Overall, the data suggest that SHP-2 is dispensable for Ag-specific CD8⁺ T cell formation. Interestingly, preliminary analysis of SHP-1^fl/fl-CD4-Cre conditional mice indicates that there is no gross defect in T cell and iNKT cell development (Figure S3 in Supplementary Material), suggesting a possible redundancy between SHP-1 and SHP-2 in T cell development. Generation of mice conditionally deficient for both SHP-1 and SHP-2 in the T cell lineage will clarify this issue.

Since SHP-2 is a regulator of lymphocyte activation, we hypothesized that aging SHP-2-deficient mice might be predisposed to autoimmune disease, as reported for other T cell tyrosine protein phosphatases (33–35). Although we did not observe any signs of autoimmunity (data not shown), we surprisingly found that 6 month-old SHP-2^fl/fl-CD4-Cre mice exhibited markedly decreased mobility, while littermate controls did not. Faxitron X-ray radiographs and μ-CT of SHP-2^−/− mice revealed enlarged bony masses at the wrists (Figures 4A,B) and knees (data not shown). Moreover, X-rays showed kinky tails (Figure S4A in Supplementary Material) and spinal abnormalities (Figure S4B in Supplementary Material) in most SHP-2^−/− animals, likely due to overgrowth of the bones in the vertebral column (data not shown). We analyzed a larger group of mice (>30 WT and 30 SHP-2^−/−) and found this phenotype was gender independent and exhibited 100% penetrance in SHP-2-deficient animals at 5 months of age. Faxitron radiographs indicated that the enlarged bony masses at the wrists were not detectable before 4 months of age (data not shown). H&E staining of the distal radius showed significant differences in bone morphology and revealed that SHP-2^−/− mice appear to have cartilage tumors resembling miniature growth plates (Figure 4C). The wrist tumor consists of large areas of cartilage tissue (1–3), indicated by positive Safranin O staining (Figure 4D). Based on their morphology and staining, the cells in these cartilaginous regions appear to be a combination of large chondrocyte-like cells and fibrochondrocyte-like cells, which is not observed in the heterozygote littermate controls. Normal bone growth and development rely on coordination between chondrocyte mass regulation and cartilage structure establishment. Overall, the phenotype observed is reminiscent of metachondromatosis. To this end, both heterozygous Ptpn11 frame shift mutations and deletions in chondroid cells have been shown to cause metachondromatosis (36), suggesting that Ptpn11 is a tumor suppressor gene in cartilage.

The previous experiments utilized SHP-2^−/− mice in which floxed Ptpn11 was deleted by Cre recombinase under the control of the cd4 promoter/enhancer/silencer. However, other mammalian cell lineages, including lymphoid tissue-inducer (LTi) cells, have also been reported to express CD4 during development and the observed phenotype could potentially be explained by CD4 expression outside of the T cell compartment (37, 38). In support of this hypothesis, SHP-2^fl/fl-Lck-Cre mice did not develop cartilage tumors suggesting that the phenotype observed was T cell independent (data not shown). We then bred SHP-2^fl/fl-CD4-Cre mice on the RAG^−/− background. In the resultant SHP-2^fl/fl-CD4-Cre-RAG^−/− mice, cartilage tumors developed (Figure S5 in Supplementary Material). The cartilage tumors in these mice can be detected at approximately the same age as the Shp-2^fl/fl-CD4-Cre mice. Because SHP-2^fl/fl-CD4-Cre-RAG^−/− mice do not have T cells, we concluded that the phenotype observed is T cell independent. Altogether, these data demonstrate that Ptpn11 gene deletion in a non-T cell subset leads to cartilage tumors.

To mark cells that have deleted the Ptpn11 gene, we crossed the SHP-2^fl/fl-CD4-Cre mice to RosaYFP reporter mice (39). In the resultant SHP-2^fl/fl-CD4-Cre-RosaYFP mice, Cre activity simultaneously deletes the ptpn11 gene locus and removes a loxP-flanked stop cassette to drive constitutive expression of a YFP gene. This fate mapping strategy allowed us to identify SHP-2-deficient cells even if they no longer express CD4. Using these mice, we found that a low number of CD45⁺CD19⁻CD3⁻TCRβ⁻ cells are YFP positive in spleens (Figure 5A), lymph nodes (data not shown), and liver (data not shown) of SHP-2^fl/fl-CD4-Cre-RosaYFP mice. The proportion of these cells are low and appear to derive from different lineages. For instance, in the spleen a subset of the CD45⁺CD3⁻ CD19⁺ cells (not shown) and a subset of the CD45⁺CD3⁻ NK1.1⁺ cells express YFP (Figure 5A). In addition, within the YFP⁺ population, a CD3⁻CD19⁻NKp46⁻NK1.1⁻CD127⁺ subset can be identified, which can be subdivided into CD4⁺ and CD4⁻ cells (Figure 5A). These cells are reminiscent of LTi-like cells (37, 38); however, only a few of these cells express CCR6 (Figure 5A).

Importantly, we have been able to detect few YFP-positive cells in frozen wrist sections of aged Shp-2^fl/fl-CD4-Cre-RosaYFP mice by confocal microscopy (Figure 5B). No YFP-positive cells were found in frozen sections from the control heterozygote mice (Figure 5B), indicating that localization of the cells of interest are concomitant to the wrist tumors. These data confirm that several non-T cell subsets express CD4 or had expressed CD4 during development. In the spleen, we have shown that SHP-2 is efficiently deleted from YFP-positive cells (Figure 1), indicating that one of these subsets could be responsible for the phenotype observed.