The serine protease TMPRSS6 is required to sense iron deficiency

Abstract

Hepcidin, a liver-derived protein that restricts enteric iron absorption, is the key regulator of body iron content. Several proteins induce expression of the hepcidin-encoding gene Hamp in response to infection or high levels of iron. However, mechanism(s) of Hamp suppression during iron depletion are poorly understood. We describe mask: a recessive, chemically induced mutant mouse phenotype, characterized by progressive loss of body (but not facial) hair and microcytic anemia. The mask phenotype results from reduced absorption of dietary iron caused by high levels of hepcidin and is due to a splicing defect in the transmembrane serine protease 6 gene Tmprss6. Overexpression of normal TMPRSS6 protein suppresses activation of the Hamp promoter, and the TMPRSS6 cytoplasmic domain mediates Hamp suppression via proximal promoter element(s). TMPRSS6 is an essential component of a pathway that detects iron deficiency and blocks Hamp transcription, permitting enhanced dietary iron absorption.

Fig. 1. The mask phenotype

A. Appearance of mask mice at 4 weeks of age. B. Anemia and iron deficiency in mask homozygotes 6-8 weeks of age. Hemoglobin (Hb), mean red blood cell volume (MCV) and plasma iron levels in mask homozygote and control mice. N=4 for each column. Error bars represent 1 standard deviation. C. and D. Retention of radioactive iron administered intragastrically. In C, mice were maintained on a normal diet prior to instillation of iron. In D, mice were iron depleted for one month prior to instillation of iron. E. After measuring retention of ⁵⁹Fe in iron-depleted mice, hepcidin mRNA levels were measured in the liver of the same five animals, normalized to mRNA levels of ribosomal protein S18. Color coding preserved for individual mice. Re-growth of hair in the mouse shown in panel A, one week (F) and two weeks (G) after initiation of a high-iron diet.

Fig. 2. The mask mutation

A. Schematic illustration of the transcripts resulting from the mutated Tmprss6^mask allele. The mutation site, at the acceptor splice junction upstream from exon 15, is indicated by black arrowhead; a premature stop codon is predicted as indicated by the asterisk. B. Illustration of the coding change that results from the mutation (SMART diagram).

Fig. 3. Rescue of the mask phenotype by BAC transgenesis

Four founders (designated 49-52 OQ) were produced by microinjecting a BAC clone bearing the WT Tmprss6 sequence into single-cell embryos homozygous for the mask mutation. Hemoglobin and serum iron levels in each of the founders, in blood sampled at 4 weeks of age. B6 controls, age-matched C57BL/6J mice, N=4. Non-transgenic controls, Tmprss6^msk/msk littermates that lacked the transgene, N=4. Error bars indicate 1 SEM.

Fig. 4. Effect of WT or mask mutant TMPRSS6 expression on Hamp promoter responses

HepG2 cells were transfected with a full-length Hamp promoter reporter construct in all experiments shown in A-D, and with a normalizing renilla luciferase construct. Tmprss6, full length construct encoding the native protein. Tmprss6^msk, the truncated mutant sequence expressed by mask mice. Tmprss6^S762A, protease-dead Tmprss6 mutant. delN, GFP cytoplasmic domain, TMPRSS6 transmembrane and ectodomain. delC, GFP ectodomain and TMPRSS6 transmembrane and cytoplasmic domains. A. The effect of Tmprss6 and Tmprss6^msk on Hamp promoter activity in cells cotransfected with a hemojuvelin-encoding vector or stimulated with BMP2, BMP4, BMP9, IL-6, or IL-1α. B. Comparison of the Hamp promoter-suppressing effects of Tmprss6, Tmprss6^msk, and Tmprss6^S762A in cells subjected to the stimuli indicated. C. Localization of delN and delC constructs in HepG2 cells by confocal fluorescence microscopy. D. Comparison of the Hamp promoter-suppressing effects of Tmprss6, Tmprss6^msk, delN and delC constructs in cells subjected to the stimuli indicated. E. Response of Hamp promoter deletion constructs to induction by IL-6 (length indicated with respect to the cap site), and the relative suppression of responses by co-transfection with Tmprss6 or Tmprss6^msk constructs. In all experiments (A-E), duplicate transfections were performed. Incubations with cytokines were performed for 16 hours before luminescence was read to estimate hepcidin promoter activity. Error bars indicate 1 SEM.