Comments for:
Abstract
Ion channel-controlled cell volume regulation is of fundamental significance to the physiological function of sperm. In addition to volume regulation, LRRC8A-dependent volume-regulated anion channel (VRAC) activity is involved in cell cycle progression, insulin signaling, and cisplatin resistance. Nevertheless, the contribution of LRRC8A and its dependent VRAC activity in the germ cell lineage remain unknown. By utilizing a spontaneous Lrrc8a mouse mutation (c.1325delTG, p.F443*) and genetically engineered mouse models, we demonstrate that LRRC8A-dependent VRAC activity is essential for male germ cell development and fertility. Lrrc8a-null male germ cells undergo progressive degeneration independent of the apoptotic pathway during postnatal testicular development. Lrrc8a-deficient mouse sperm exhibit multiple morphological abnormalities of the flagella (MMAF), a feature commonly observed in the sperm of infertile human patients. Importantly, we identified a human patient with a rare LRRC8A hypomorphic mutation (c.1634G>A, p.Arg545His) possibly linked to Sertoli cell–only syndrome (SCOS), a male sterility disorder characterized by the loss of germ cells. Thus, LRRC8A is a critical factor required for germ cell development and volume regulation in the mouse, and it might serve as a novel diagnostic and therapeutic target for SCOS patients.
Authors
Jianqiang Bao, Carlos J. Perez, Jeesun Kim, Huan Zhang, Caitlin J. Murphy, Tewfik Hamidi, Jean Jaubert, Craig D. Platt, Janet Chou, Meichun Deng, Meng-Hua Zhou, Yuying Huang, Héctor Gaitán-Peñas, Jean-Louis Guénet, Kevin Lin, Yue Lu, Taiping Chen, Mark T. Bedford, Sharon Y.R. Dent, John H. Richburg, Raúl Estévez, Hui-Lin Pan, Raif S. Geha, Qinghua Shi, Fernando Benavides
No evidence for a role of mutations affecting LRRC8/VRAC channels in human infertility
Submitter: Thomas J. Jentsch | jentsch@mdc-berlin.de
Authors: Thomas J. Jentsch, Jennifer C. Lück, and Florian Ullrich
Leibniz-Forschungsinstitut für Molekulare Pharmakologie (FMP) & Max-Delbrück-Centrum für Molekulare Medizin (MDC)
Published October 2, 2018
In a recent issue of JCI Insight Bao et al. (Bao et al, 2018) reported on male infertility in mice and humans related to mutations in the LRRC8A subunit of the volume-regulated anion channel (VRAC) (Jentsch, 2016). The authors’ analysis of the germ cell phenotype of Lrrc8a mutant mice appears solid, mostly well done, and largely agrees with our work (Lück et al, 2018). However, their suggestion that a pArg545His LRRC8A variant causes male infertility in humans (specifically Sertoli cell only syndrome, SCOS), an important part of the paper which is prominently mentioned in the abstract, is unfounded. Indeed, their own functional analysis (Bao et al, 2018) rather demonstrates that this sequence variant is almost certainly a benign polymorphism. There is no evidence that LRRC8A mutations underlie human non-syndromic infertility and the publication by Bao et al. is grossly misleading in this respect.
Bao et al. identified an LRRC8A sequence variant on one allele of a single patient with SCOS, a non-syndromic (i.e. not associated with other symptoms) male infertility disorder. Since the identification of a sequence variant in a single patient cannot prove that the variant causes the disease, the authors try to bolster this notion using functional analysis of the variant in heterologous expression. They report that the LRRC8A variant causes a 25-30% decrease of anion current amplitude when co-expressed 1:1 with another channel subunit, either LRRC8C or LRRC8D (VRACs are composed of the obligatory LRRC8A subunit and any of LRRC8B to LRRC8D (Voss et al, 2014) and assemble to hexamers (Deneka et al, 2018; Kefauver et al, 2018)). They speculate, but do not show, that in a ‘real’ situation the loss of current amounts to more than 60% (Bao et al, 2018). Unfortunately, their analysis and conclusions are fundamentally flawed.
First, they used an artificial expression system in which – by an unknown mechanism - the addition of GFP variants to the C-termini of both subunits leads to currents under isotonic conditions (Gaitán-Peñas et al, 2018), obviating the need for cell swelling that is normally required to activate VRAC. Whereas this trick is convenient for studying channel pore properties, it is unsuited to study effects of point mutations on current amplitudes because these effects depend on channel activation. This activation, however, is fundamentally changed by the C-terminal fusion.
Second, even if real, a 25-30% reduction of current amplitudes from LRRC8A/C or LRRC8A/D channels is small and unlikely to cause disease. This is because heterozygous Lrrc8a+/- mice are viable, fertile, and lack any apparent phenotype (Kumar et al, 2014). It appears that VRAC currents must be reduced way below 50% to cause infertility like in the spontaneous mouse mutant ébouriffé (Lalouette et al, 1996; Platt et al, 2017) or in Lrrc8-/- mice that completely lack VRAC (Kumar et al, 2014). However, such a strong reduction is associated with additional, often severe symptoms (Kumar et al, 2014; Lalouette et al, 1996; Platt et al, 2017), that is, it causes syndromic infertility.
Third, and most importantly, the authors themselves showed that the pArg545His variant is a benign polymorphism, although they eschew this conclusion. Because this variant is present in a heterozygous state in the SCOS patient, Benavides and coworkers (Bao et al, 2018) expressed 50% WT and 50% variant LRRC8A together with either LRRC8C or LRRC8D (always fused to GFP). Now they observed no reduction in current amplitudes. Hence the mutation has no functional consequences under somewhat more realistic conditions.
Fourth, based on the ~30% reduction of current amplitudes from LRRC8A/C or LRRC8A/D channels by the LRRC8A variant (in a ‘homozygous state’), the authors state that ‘it is reasonable to speculate that the true reduction of VRAC currents due to the LRRC8AR545H mutation in a single cell could exceed 60%’. This is not ‘reasonable’, but obviously nonsense: a mixture of these channels is expected to lead again to a ~30% reduction.
Since native cells express several LRRC8 isoforms and single LRRC8/VRAC channels can contain at least three different LRRC8 proteins (Lutter et al, 2017), the reduction in VRAC currents in native cells is rather expected to be less, as convincingly shown by the authors’ experiments emulating a heterozygous mutation. Nonetheless, the authors suggest that the LRRC8A variant may have stronger effects in germ cells which express LRRC8A – LRRC8D. However, they do not test this idea by co-expressing variant LRRC8A with WT LRRC8A, LRRC8B, -C and –D in their oocyte system. Their argument that they could not do this experiment because co-expression of LRRC8A with only LRRC8B does not yield currents does not make sense. Native cells do express these combinations and the incorporation of LRRC8B into channels containing LRRC8A and a third LRRC8 isoform influences VRAC transport properties (Lutter et al, 2017).
At this point it might no longer be necessary to point out that the homologous LRRC8D subunit displays a histidine at the equivalent (cytoplasmic) position that is changed to histidine in the reported LRRC8A variant.
In conclusion, there is no evidence that LRRC8A mutations cause human infertility, but compelling evidence that the LRRC8A sequence variant described by Bao et al. is a benign polymorphism. The asserted link between LRRC8A and human infertility (Bao et al, 2018) is highly misleading and will confuse the field.
References
Response to Jentsch et al.
Submitter: Fernando Benavides | fbenavid@mdanderson.org
Authors: Raúl Estévez and Fernando Benavides
MD Anderson
Published October 2, 2018
We thank Dr. Jentsch and colleagues for their interest in our work and agree with them that we do not provide definitive evidence that the pArg545His variant of LRRC8A causes Sertoli cell-only syndrome in human males. Our intention was to define the role of cell volume regulation in germ cells using a mouse model. Therefore, when we stated in our article abstract that "we identified a rare hypomorphic mutation possibly linked to Sertoli cell-only syndrome (SCOS),” we were careful to avoid stating that we identified a definitive link between the LRRC8AR545H mutation and human SCOS. Further, we avoided using the word “human” in the title of the paper. For these reasons, we disagree with the conclusion of Jentsch et al. that our work will confuse the field and that our suggestion of a possible link between LRRC8A and SCOS is “grossly misleading.”
In response to the first point of Jentsch et al. regarding the use of Xenopus oocytes to characterize the mutation, this system has been long-established to study ion channels generally, and more recently, LRRC8 subunits specifically. Although it is an artificial system: a) the fusion of GFP to the C terminus does not abolish VRAC channel activation by swelling; and b) all of the properties reported for LRRC8 proteins in cells have been reproduced in this system. The expression system preferred by Jentsch et al. is using LRRC8 knock-out cells that overexpress LRRC8 proteins. This system is also less than ideal. Indeed, the authors of the letter have reported that the properties of channels created using this overexpression system are very different than endogenously expressed channels (Ullrich F et al, 2016). These altered properties are likely due to changes in channel component stoichiometry, as was demonstrated in the oocyte system when components are overexpressed (Gaitán-Peñas et al, 2018). Thus, the only way to properly study this mutation will be to develop a knock-in mouse or cell line model, as we proposed in the discussion section of our paper (Bao et al, 2018).
In response to the second point, regarding what Jentsch et al. regard as modest changes in current in our model, we would point out that one cannot always extrapolate mouse data to predict human outcomes as mouse models do not always perfectly recapitulate human phenotypes. A modest effect in our model might have a more profound effect in other systems. For example, the authors’ own mouse models of leukodystrophies (Hoegg-Beiler MB 2014) do not fully capture the human disease. Thus, it may be that human stem cells (like those of the blood and germ lines) may be more sensitive than their mouse counterparts such that a 25-30% reduction in current amplitude may result in a phenotype. Nonetheless, although our genetic evidence is strong, we cannot be certain that this polymorphism causes the effect, as the effect is small. For this reason, we stated in the abstract only that the polymorphism may be “possibly linked” to SCOS.
In response to the third point, regarding the interpretation by Jentsch et al. that the pArg545His variant is a benign polymorphism, we believe that they have misunderstood our conclusions. In co-expression experiments (50% WT + 50% mutant), we observed a 15% reduction in the current. Thus, we indicate that the polymorphism is not a dominant mutation. It does not have a dominant effect over the other subunit. For this reason, we concluded that pArg545His is a “potentially” hypomorphic mutation.
In response to the fourth point, again regarding reduction in current amplitudes, we agree with the criticism that the speculation about current amplitudes and different subunit combinations was not judicious; we are unsure of the precise stoichiometry of the subunits in our channels. Our point was merely that if the total number of LRRC8A mutant subunits within the VRAC channel increases, then mutations affecting LRRC8A may have a greater effect. But, we agree with Jentsch et al. that the sentence, as written, may be overly speculative. We co-expressed the LRRC8A mutant with all the accessory subunits independently (LRRC8C, LRRC8D and LRRC8E) and in all combinations, we observed a 30% decrease in current. To carefully examine subunit combinations, which we agree should be done, we would prefer to create a knock-in mouse or cell line, which may better reflect the in vivo situation, rather than perform co-expression experiments, as Jentsch et al. suggest.
In response to the fifth point, regarding the identity of the residue at the equivalent cytoplasmic position, even if the LRRC8D subunit displays a histidine in the same position as the LRRC8A mutant, our conclusions do not change.
In summary, we have only suggested that the mutation might be a hypomorphic polymorphism that causes a modest (30%) reduction in function. It cannot be concluded, based on our data, that this is a mutation that causes SCOS in humans, and we have not made this claim. We only mentioned that the mutation may possibly be linked to the human phenotype. This has been clearly indicated in the both the abstract and the article. This should be neither misleading nor confusing to those in the field.
References
3. M.B. Hoegg-Beiler, S. Sirisi, I.J. Orozco, I. Ferrer, S. Hohensee, M. Auberson, K. Gödde, C. Vilches, M. López de Heredia, V. Nunes, Estévez R, T.J. Jentsch (2014). Disrupting MLC1 and GlialCAM and ClC-2 interactions in leukodystrophy entails glial chloride channel dysfunction. Nature Communications 5: 3475.
4. Ullrich F, Reincke SM, Voss FK, Stauber T, Jentsch TJ (2016). Inactivation and anion selectivity of volume-regulated anion channels depend on c-terminal residues on the first extracellular loop. Journal of Biological Chemistry 291:17040-8.
