How to Assess Causality of TMPRSS6 Mutations?


  • Contract grant sponsors: e-Rare 2009; MIUR- PRIN (Progetto di Rilevante Interesse Nazionale) 2010–2011.

We have read with interest Guillem et al. (2012), which reports and characterizes new TMPRSS6 mutations in patients with iron refractory iron deficiency anemia (IRIDA) [Finberg et al., 2008]. The authors propose a method to assess the causality of novel mutations. To this end, the authors compare the results obtained by using a hepcidin–luciferase-based assay and the determination of the serine protease activity in the culture media of cells transfected with IRIDA-associated mutations in comparison with wild type (wt) and artificial, inactive TMPRSS6 variants. They conclude that the serine protease-based test, which uses a chromogenic substrate, is a simple and preferential method to assess the pathogenic role of novel mutations.

The authors showed that in the luciferase assay some IRIDA-associated mutants (E114K; L235P; Y418C and P765A) retain the ability to suppress hepcidin transcription. In addition, the artificial variants R576A and S762A, expected to be inactive as they affect the autocatalytic cleavage site and the protease catalytic triad, respectively, also inhibit hepcidin (see Fig. 3 of Guillem et al., 2012]. For this reason, the authors concluded that the hepcidin–luciferase-based assay is inappropriate to evaluate the functionality of TMPRSS6 mutants. We were surprised by this conclusion, as we have extensively used a similar assay to characterize TMPRSS6 mutations identified in IRIDA patients [De Falco et al., 2010; Silvestri et al., 2008; 2009]. In our hands, all the mutations tested showed reduced inhibitory activities, except R271Q, that in vivo is associated with the I212T causal mutation [De Falco et al., 2010]. Other defective TMPRSS6 variants were characterized by this assay [Altamura et al., 2010; Ramsay et al., 2009]. Furthermore, using this approach, we were able to detect the small difference in the inhibitory activity of the common TMPRSS6 polymorphic variant (rs855791 or V736A) and to show that the difference was well correlated to serum hepcidin level variations in normal subjects [Nai et al., 2011].

We had the opportunity to analyze the new TMPRSS6 variants described by Guillem et al. (2012), which were kindly released to us by the authors. Using the Hep3B cells (Fig. 1A) and the controlled conditions of the transfection assay as previously described [Pagani et al., 2008; Silvestri et al., 2008], all the variants studied had significantly impaired ability to repress the HJV-dependent hepcidin activation (Fig. 1A). All mutants were less inactive compared with the truncated MASK variant [Du et al., 2008]. Their inhibition ability was significantly lower than that of wt and similar to the inhibition shown by the well-characterized serine protease R774C mutant [Silvestri et al., 2008; 2009]. The only exception was L235P, whose inhibitory activity was still decreased compared with the wt isoform, although the difference was not statistically significant. Interestingly, heterozygous L235P was found in a patient in combination with Y418C causal variant [Guillem et al., 2012].

Figure 1.

Functional characterization of the TMPRSS6 variants. A firefly luciferase reporter driven by 2.9 kbp (250 ng) of the proximal hepcidin promoter was cotransfected into Hep3B cells (A) and HuH7 cells (B) (48 multiwells plate) with pRL-TK (15 ng), either alone or with HJV (50 ng) and/or TMPRSS6 (10 ng) expressing vectors. Relative luciferase activity is calculated as a ratio between firely and renilla and expressed as a multiple of the activity of cells transfected with the reporter alone. Error bars indicate standard error. C: HeLa cells were transfected with HJV wt in the presence of the empty vector (mock), TMPRSS6 wt (wt), TMPRSS6 R576A (R576A), TMPRSS6 S762A (S762A), and TMPRSS6 MASK (MASK). Concentrated media and membrane proteins were loaded onto a 10% SDS-PAGE and processed for western blot analysis. Anti-FLAG and anti-HJV were used to detect TMPRSS6 and HJV, respectively. Anti-panCadherin was used to detect panCadherin. D: Binding assay was used to measure membrane-HJV (m-HJV) in the presence of TMPRSS6 wt, R576A, S762A, and MASK. HeLa cells were transiently transfected with the wt and mutant expressing vectors, or the empty vector. The amount of m-HJV was calculated as the ratio between the absorbance of unpermeabilized and permeabilized cells. Error bars indicate standard error. E: Characterization of E114K and L235P TMPRSS6 variants. Concentrated media of transiently transfected HeLa cells were analyzed by 10% SDS-PAGE. The soluble serine protease domain of TMPRSS6 was revealed using the anti-FLAG antibody. *: unspecific band. The unpaired two-tailed Student t-test was used to analyze significant changes (GraphPad Prism Version 5.0a) compared with wt TMPRSS6. P values <0.05 were considered statistically significant. *: P < 0.05; §: P < 0.01; #: P < 0.001.

Consistent with previous results [Guillem et al., 2012], R576A and S762A inhibited hepcidin activation even in our assay (Fig. 1A).

Using the HuH7 cells, that are hemochromatosis cell lines [Vecchi et al., 2010], a trend toward an impaired inhibitory activity of MT2 mutants in the hepcidin promoter luciferase-based assay was observed (Fig. 1B), although no statistical significance was reached for almost all the MT2 mutants analyzed, with the exception of Y418C and the MASK variant.

As TMPRSS6 cleaves membrane-HJV [Silvestri et al., 2008], we then moved to study the HJV proteolytic cleavage. R576A and S762A were both unable to cleave HJV, as they did not release soluble fragments in the medium (Fig. 1C, upper panel). In accordance, they did not release the serine protease domain, confirming that they are proteolytically inactive. These results were confirmed using the HJVR335Q variant (data not shown), which does not release soluble HJV but localizes to the cell surface as the wt isoform [Silvestri et al., 2008].

To investigate why proteolytically inactive variants suppress hepcidin transcription in the promoter assay, we analyzed the plasma membrane localization of TMPRSS6 artificial variants and HJV by using both cell surface biotinylation and binding assay [Silvestri et al., 2007]. Both TMPRSS6 variants are expressed on the cell surface as the wt protein (Fig. 1C, lower panel). When coexpressed with the proteolytically inactive MASK, HJV is found in large proportion on cell membrane (Fig. 1C, panel “m-HJV” and Fig. 1D), whereas the amount of membrane HJV in coexpression with R576A and S762A TMPRSS6 variants was highly reduced at the level reached by the wt serine protease (Fig. 1C, panel “m-HJV” and Fig. 1D). The same results were confirmed in HuH7 cells (data not shown). Because this reduction cannot be ascribed to HJV cleavage, we suggest that the two artificial mutants interfere with the HJV plasma membrane localization, although the mechanism remains to be clarified. These results may explain the lower degree of hepcidin activation of R576A and S762A compared with the truncated variant MASK observed in Figure 1A and B.

Overall, we suggest a note of caution in the interpretation of the data obtained by the luciferase promoter assay when using insufficiently characterized TMPRSS6 variants. Under controlled experimental conditions (i.e., low concentrations of HJV and TMPRSS6 expressing vectors) and the appropriate (Hep3B) cell line the hepcidin–luciferase-based method seems to be more sensitive compared with the proposed biochemical assay. The proposed serine protease chromogenic assay may have drawbacks. Small but significant amounts of serine protease domain are released from E114K and L235P variants (Figure 1E); however, no activity was detected by the biochemical assay and the cut off of sensitivity of this method remains to be defined. Our conclusion is that neither the chromogenic method nor the hepcidin–luciferase-based assay is suitable for routine studies.

An ideal simple method to assess the function of TMPRSS6 variants has not likely been identified. Considering that IRIDA patients have normal/high serum hepcidin levels despite the condition of iron deficiency that is usually associated with low/undetectable hepcidin [Traglia et al., 2011], if assays to measure serum hepcidin become available in clinics, they will likely substitute the in vitro costly and time consuming study of TMPRSS6 mutants [Camaschella, 2013].


We thank Flavia Guillem and Bernard Grandchamp for the gift of TMPRSS6 mutant constructs.

Disclosure statement: The authors declare no conflict of interest.