Deletion of Trpm7 Disrupts Embryonic Development and Thymopoiesis Without Altering Mg2+ Homeostasis

See allHide authors and affiliations

Science  31 Oct 2008:
Vol. 322, Issue 5902, pp. 756-760
DOI: 10.1126/science.1163493


The gene transient receptor potential-melastatin-like 7 (Trpm7) encodes a protein that functions as an ion channel and a kinase. TRPM7 has been proposed to be required for cellular Mg2+ homeostasis in vertebrates. Deletion of mouse Trpm7 revealed that it is essential for embryonic development. Tissue-specific deletion of Trpm7 in the T cell lineage disrupted thymopoiesis, which led to a developmental block of thymocytes at the double-negative stage and a progressive depletion of thymic medullary cells. However, deletion of Trpm7 in T cells did not affect acute uptake of Mg2+ or the maintenance of total cellular Mg2+. Trpm7-deficient thymocytes exhibited dysregulated synthesis of many growth factors that are necessary for the differentiation and maintenance of thymic epithelial cells. The thymic medullary cells lost signal transducer and activator of transcription 3 activity, which accounts for their depletion when Trpm7 is disrupted in thymocytes.

The transient receptor potential (TRP) superfamily comprises cation-permeant ion channels that have diverse functions (13). TRPM7 (1, 2) and TRPM6 (4, 5) proteins also contain a C-terminal kinase domain (6). TRPM7 is expressed in all examined cell types (3) and mediates the outwardly rectifying Mg2+-inhibitable current (MIC) (7). TRPM6 and TRPM7 exhibit nearly identical current-voltage (I-V) relations, conducting only a few pA of inward current at physiological pH levels (1, 2, 8, 9).

A chicken DT-40 B cell line targeted for Trpm7 gene disruption was reported to require high concentrations of extracellular Mg2+ (10 mM) for survival (10). Given the permeability of TRPM7 to Mg2+, the results have been interpreted to indicate that TRPM7 was critical for cellular Mg2+ homeostasis in vertebrates. A role for TRPM7 in vertebrate development was suggested by a Danio rerio Trpm7 mutant that exhibited abnormal skeletogenesis and melanophore development, but whether this developmental defect is related to Mg2+ homeostasis remains unclear (11).

We generated multiple mouse lines with a targeted deletion of the Trpm7 gene (fig. S1A) (12). Mouse lines with disruption of Trpm7 in all tissues (global deletion), generated using three different approaches, did not yield any live Trpm7null/null animals. Mendelian ratios of littermate genotypes indicated that Trpm7null/null mice died prenatally (194 pups analyzed) (fig. S1, D and E). To rule out the possibility that embryonic viability was compromised by the disruption of maternoembryonic transport, we deleted Trpm7 by using paternal Sox-2 Cre, which deletes the gene in the embryonic cells but not in extraembryonic visceral endoderm (13). Eight litters from this mating scheme failed to produce Trpm7–/fl (Sox-2 Cre) mice, indicating that embryonic lethality resulted from a requirement for TRPM7 in the developing embryo rather than a compromise of maternoembryonic nutrient transport. For further analysis, we used mice in which a β-geo cassette (coding for β-galactosidase) was inserted in the first intron of Trpm7 (Trpm7geo) (12). The β-geo transcript has a splice acceptor site but not a splice donor site; thus, barring alternative splicing, Trpm7geo generates a null allele of Trpm7. We isolated and analyzed embryos from Trpm7geo/+ intercrosses at various times after fertilization and determined that Trpm7null/null embryos died before day 7.5 of embryogenesis (E7.5) (fig. S1F). LacZ staining of Trpm7geo/+ embryos revealed a predominant expression in the fetal heart at E9.5, followed by a gradual and intense expansion of the expression across the ventral region at E10.5, peaking throughout the embryo at E11.5 and E12.5. The broad expression pattern was maintained through E14.5 (fig. S1G). Thus, TRPM7 is expressed in embryonic stem cells (fig. S1B), expression is increased in the early embryo, and the expressed TRPM7 has a nonredundant and vital role in the embryonic development of the mouse.

Using lck-Cre mice, we selectively deleted Trpm7 in developing thymocytes. Deletions of Trpm7–exon 17, in thymocytes and mature T-lymphocytes isolated from Trpm7-/fl (lck-Cre) mice, were confirmed by reverse transcription polymerase chain reaction (RT-PCR) (fig. S2A) and quantified using quantitative RT-PCR directed against exon 17. Unmodified transcripts in thymocytes (8.9 ± 0.28% of normal) were lower than in T lymphocytes (16.8 ± 0.96% of normal) when compared with those in wild-type (wt) cells, which probably reflected the presence of contaminating cells of non–T cell lineage in Thy-1.2–directed immunoaffinity preparations from mouse spleens.

Whole-cell patch-clamp recordings revealed that MIC current (IMIC) in thymocytes was potentiated by extracellular application of 10 mM NH4Cl (Fig. 1A) and reversibly inhibited by 10 mM MgCl2 (Fig. 1A) or 100 μM 2-aminoethoxydiphenylborate (2-APB) (fig. S3A) (7, 14). IMIC was abrogated in T cells derived from Trpm7–/fl (lck-Cre) mice (Fig. 1, B and C), whereas K+ currents (Kv1.3, Kv3.1, and KCa 3.1) (1518) were unaffected (fig. S3B). IMIC was still present in a small population of T cells, probably because of incomplete Cre expression and inadvertent patch clamping of Thy-1.2+ splenocytes other than T lymphocytes.

Fig. 1.

IMIC and Mg2+ homeostasis in Trpm7-deficient cells. (A) I-V relationship of IMIC in wt and Trpm7-deficient [knockout (KO)] thymocytes. (B) IMIC densities in wt (n = 13 cells) and KO (n = 18 cells) T lymphocytes (P <0.0001, two samples independent t test). Box charts are shown as a box (25 to 75 percentile), vertical bars (5 to 95 percentile), and data points (diamonds) overlap with the mean value (empty square) and median value (horizontal line in the box). (C) IMIC densities in wt (n = 9 cells) and KO (n = 10 cells) thymocytes (P = <0.0001, two samples independent t test). (D) Mg2+ uptake in wt T lymphocytes loaded with KMG104AM, as indicated by the averaged ratio of fluorescence intensity F at indicated time (seconds) over the initial fluorescence F0 at 0 s. Mg2+ uptake in the absence (blue, n = 25 cells) or presence (red, n = 25 cells) of 0.5 mM 2-APB is shown. (E) ICP-MS quantitation of total Ca2+ and Mg2+ in HNO3 extracts. Average concentration of total cellular Mg2+ (n = 3 mice) as calculated by normalizing to a [K+] of 120 mM is shown. Error bars indicate ± SD. (F) Mg2+ uptake in wt (blue, n = 117 cells) and KO (red, n = 102 cells) thymocytes. [Ca2+] and [Mg2+] are in mM.

We used a cell-permeable fluorescent indicator [KMG104AM (19)] to evaluate Mg2+ influx in freshly isolated T cells from wt and Trpm7–/fl (lck-Cre) mice. T lymphocytes incubated with KMG104AM and maintained in a Mg2+-free medium responded with an increased fluorescence intensity to extracellular perfusion with solutions containing 10 mM MgCl2 but not to solutions containing 10 mM CaCl2 (fig. S3C). Mg2+ influx in wt T lymphocytes was insensitive to 0.5 mM 2-APB (Fig. 1D). Similarly, intracellular alkalinization induced by extracellular 50 mM NH4Cl, which potentiates ITRPM7 (14), did not result in higher Mg2+ influx (fig. S3D). Mg2+ influx in thymocytes freshly isolated from Trpm7-/fl (lck-Cre) mice was insensitive to deletion of TRPM7 (Fig. 1F), which indicates that TRPM7 does not mediate the observed Mg2+ influx in T cells. To test the tissue Mg2+-dependence on TRPM7, we used inductively coupled plasma mass spectrometry (ICP-MS) to measure total Mg2+ in freshly isolated T cells. The total [Mg2+] in T cells obtained from wt and Trpm7-/fl (lck-Cre) mice were not statistically different (Fig. 1E). These data indicate that TRPM7 is not essential for cellular Mg2+ homeostasis in mice.

In the intestine of adult Trpm7–/fl (lck-Cre) mice, T lymphocytes were readily detected at a density comparable to that of wt mice (fig. S7), whereas a small reduction in T cell density was evident in the lymph nodes (fig. S8) and spleen (fig. S4B) of Trpm7–/fl (lck-Cre) mice. Flow cytometry of splenocytes isolated from wt and Trpm7–/fl (lck-Cre) mice revealed a small reduction in the percentage and numbers of T cells but not of B cells (Fig. S4, C and D). Despite the decrease in the splenic T cell numbers, the results show that mature T lymphocytes in Trpm7–/fl (lck-Cre) mice are able to survive and populate the periphery.

The thymi in Trpm7–/fl (lck-Cre) mice developed morphological abnormalities with an 85% phenotypic penetrance (n = 27 mice), which suggests defective thymopoiesis. Histology of thymic sections derived from 12-week-old Trpm7–/fl (lck-Cre) and Trpm7+/fl littermate controls showed abnormal thymic architecture in Trpm7–/fl (lck-Cre) mice (Fig. 2A). The boundary between cortical and medullary areas was easily visible in wt thymi but not in Trpm7-deficient thymi (Fig. 2A and fig. S6A). In contrast to wt thymi, where the CD3+ T cells remained confined to the medulla (outlined and marked as T), the CD3+ cells in the thymi of Trpm7-/fl (lck-Cre) mice were uniformly distributed across the thymic stroma (Fig. 2B and fig. S6B). Evaluation of thymic cellularity indicated a substantial reduction in the number of thymocytes in Trpm7–/fl (lck-Cre) mice (Fig. 2C).

Fig. 2.

Deletion of Trpm7 in thymocytes leads to defective thymopoiesis. (A) Hematoxylin and eosin–staining of thymus sections from 12-week-old wt (top) and KO (bottom) mice at 4× (left) and 20× (right) magnification (red box). The boundary between medullary and cortical regions is highlighted with a solid white line where evident. (B) Thymocytes were immunolocalized by antibody-to-CD3 staining (brown) against a nuclear counterstain (blue) in the thymus sections obtained from wt (top) and Trpm7-deficient (bottom) mice. (Left) 4× magnification. Red boxes indicate the areas that are shown at 20× magnification to the right. The CD3+ T cell–enriched medullary regions are highlighted in a wt thymus (see fig. S6 for larger images). (C) Box chart showing the reduced number of thymocytes in Trpm7-deficient mice (red, n = 9 mice) as compared with wt mice (black, n = 9 mice). Box charts shown as a box (±SD), vertical bars (maximum-minimum values), and data overlap. The P values in all of the box charts were calculated using the two-sample independent t test. (D) Flow cytometry of CD4 and CD8 on thymocytes from wt and KO mice. (E) Box charts comparing the total number of thymocytes in the DN, DP, CD4+, and CD8+ thymocytes are shown (n = 7 mice).

Thymocytes from Trpm7–/fl (lck-Cre) mice contained a higher percentage (Fig. 2D) and number (Fig. 2E) of CD4– CD8– [double negative (DN)] cells than did thymocytes from wt controls, which indicates a partial developmental block in transition from the DN to double-positive (DP) stage. This developmental defect may account for the reduced number of T cells in Trpm7–/fl (lck Cre) mice. Analysis of the DN population based on the cell-surface expression of CD44 and CD25 revealed a significantly higher percentage of DN thymocytes in the DN3 (CD44– CD25+) stage (Fig. 3A), which indicates a failure to down-regulate CD25 expression during T cell development. Because of the block at the DN stage, the cell number is significantly higher in Trpm7-deficient thymi. Overlays of CD25 expression of the total thymocyte population and of DN thymocytes show that the proportion of cells expressing CD25 was significantly higher in Trpm7-deficient thymi (Fig. 3B). We calculated the changes in percentages (Fig. 3C) and number (Fig. 3D) of DN cells in DN1 (CD44+ CD25–), DN2 (CD44+ CD25+), DN3 (CD44– CD25+), and DN4 (CD44– CD25–) stages. These data indicated that a portion of Trpm7-deficient thymocytes fails to down-regulate high-affinity interleukin-2 receptors (CD25), exhibiting a block during the transition from the DN3 to DN4 stage. Cell-surface expression of T cell receptor β (TCR-β) chain (fig. S5C) was not substantially altered, which suggests that the developmental defect was not due to a failure in TCR-β locus rearrangement.

Fig. 3.

Trpm7-deficient thymocytes are partially blocked at the DN3 stage. (A) Flow cytometry of CD44 and CD25 expression in DN (CD4– CD8–) thymocytes. Thymocytes were stained with antibody to CD4 [fluorescein isothiocyanate (FITC)], antibody to CD8 (FITC), antibody to CD44 [phycoerythrin (PE)] and antibody to CD25 [phosphatidylcholine 7 (PC7)]. FITC-negative cells (DN) were analyzed for CD44 and CD25 expression. (B) (Left) Overlay of cell-surface CD25 expression in wt (black) and Trpm7-deficient (red) thymocytes. (Right) Overlay of CD25 expression in DN thymocytes. (C) Box charts showing percentage of DN population found in the (left to right) DN1, DN2, DN3, and DN4 stages. (D) Box charts showing total number of thymocytes found in the (left to right) DN1, DN2, DN3, and DN4 stages.

We found a progressive loss of thymic medullary cells [cytokeratin 5+ (K5+); Fig. 4A, green] but not thymic cortical cells (K8+; Fig. 4A, red) in comparing 4- and 12-week-old wt and Trpm7–/fl (lck-Cre) mouse thymic sections. In wt mice, the CD3+ cells (Fig. 4B, green) were distributed preferentially within medullary regions of the thymus and showed minimal overlap with cortical thymic epithelial cells (TECs) (K8+; Fig. 4B, red). In contrast, the loss of medullary regions in Trpm7-deficient mice was accompanied by a uniform distribution of CD3+ thymocytes in the thymic cortex, as detected by an extensive overlap of K8 and CD3 staining (Fig. 4B).

Fig. 4.

Deletion of Trpm7 in thymocytes results in progressive loss of medullary epithelial cells. (A) Immunofluoresence staining of thymus sections with antibodies to thymic epithelial markers K8 (red) and K5 (green). Letters indicate thymic cortex (C) and medullary (M) regions. Scale bar, 200 μm. (B) Staining of thymus sections to CD3 (green) and K8 (red). (C) Dysregulated mRNA encoding growth factors in knockout thymocytes relative to wt thymocytes identified by quantitative RT-PCR. Growth factors with increased (blue) or decreased (red) mRNA abundance are presented as average ΔΔCt values (n = 3 mice). Error bars indicate SD. A complete list of quantitative RT-PCR results is in fig. S9. (D) Immunofluoresence staining of thymus sections with antibody to STAT3 (red) and K5 (green). (E) Immunofluoresence staining of phospho-STAT3 (Tyr705) (red) and K5 (green) in thymus sections. Scale bar, 20 μm.

We conducted a quantitative RT-PCR analysis of freshly isolated thymocytes for mRNA that encoded 82 growth factors with proposed roles in tissue growth and maintenance (fig. S9). We identified seven growth-factor mRNAs whose abundance increased by more than threefold and five growth-factor mRNAs present at <33% of normal levels in Trpm7-deficient thymocytes (Fig. 4C). Growth-factor mRNAs that were down-regulated included fibroblast growth factor 13 (FGF-13), FGF-7, and midkine. FGF-7 is an important growth factor for thymic epithelial cells (20, 21), and FGF receptors activate signal transducer and activator of transcription 3 (STAT3)–mediated transcriptional responses (22), a pathway crucial for the maintenance of thymic medullary cells (23). Midkine induces mesenchymal-epithelial transition through the activation of STAT3 (24, 25).

Because STAT3 is autoregulated, the levels of STAT3 are a useful indicator of ongoing STAT3 activity (26). In wt mice, STAT3 was specifically expressed in medullary TECs (identified by expression of the K5 marker) and progressively lost in medullary TECs in Trpm7-deficient thymi (Fig. 4D). In 12-week-old Trpm7–/fl (lck-Cre) mice, STAT3 was not detectable in the remnants of the atrophic thymic medulla. Similarly, although activated phospho-STAT3 was readily detected immunohistochemically in the nucleus of wt medullary TECs, there was no evidence of activated phospho-STAT3 in Trpm7-deficient medullary TECs (Fig. 4E). These data show that deletion of Trpm7 in thymocytes results in reduced STAT3 activity and abundance in thymic medullary cells, which is expected to lead to a progressive loss of thymic architecture.

TRPM7 is the first TRP channel to be identified with a nonredundant role in embryogenesis and the only ion channel known to be necessary for thymopoiesis. The most notable feature of TRPM7 is the permeation of Ca2+, Mg2+, and trace metals in the very same structure that contains a kinase. TRPM7 mediates exceedingly low inward conductance, which suggests that the actions of the permeant species are localized and do not substantially affect global Mg2+ levels. Our work is now concentrated on how this bifunctional protein mediates these effects on cell-differentiation processes.

Supporting Online Material

Materials and Methods

SOM Text

Figs. S1 to S9


References and Notes

Stay Connected to Science

Navigate This Article