Biochemical Journal

Research article

Aminoaciduria, but normal thyroid hormone levels and signalling, in mice lacking the amino acid and thyroid hormone transporter Slc7a8

Doreen Braun, Eva K. Wirth, Franziska Wohlgemuth, Nathalie Reix, Marc O. Klein, Annette Grüters, Josef Köhrle, Ulrich Schweizer


LAT2 (system L amino acid transporter 2) is composed of the subunits Slc7a8/Lat2 and Slc3a2/4F2hc. This transporter is highly expressed along the basolateral membranes of absorptive epithelia in kidney and small intestine, but is also abundant in the brain. Lat2 is an energy-independent exchanger of neutral amino acids, and was shown to transport thyroid hormones. We report in the present paper that targeted inactivation of Slc7a8 leads to increased urinary loss of small neutral amino acids. Development and growth of Slc7a8−/− mice appears normal, suggesting functional compensation of neutral amino acid transport by alternative transporters in kidney, intestine and placenta. Movement co-ordination is slightly impaired in mutant mice, although cerebellar development and structure remained inconspicuous. Circulating thyroid hormones, thyrotropin and thyroid hormone-responsive genes remained unchanged in Slc7a8−/− mice, possibly because of functional compensation by the thyroid hormone transporter Mct8 (monocarboxylate transporter 8), which is co-expressed in many cell types. The reason for the mild neurological phenotype remains unresolved.

  • brain
  • monocarboxylate transporter 8 (Mct8)
  • system L amino acid transporter 2 (Lat2)
  • 3,3′,5,5′-tetra-iodo-L-thyronine (T4)
  • thyroid-stimulating hormone (TSH)
  • 3,5,5′-tri-iodo-Lthyronine (T3)


Physiological studies have established that transport of amino acids across the plasma membrane requires the presence of distinct transport ‘systems’, because the zwitter-ionic compounds cannot cross the lipid bilayer by passive diffusion [1]. These ‘systems’ display broad substrate specificities and are accordingly termed neutral, basic, acidic, iminoglycine and β-amino systems [2]. Amino acid transport is not specific for intestinal and renal tubule epithelia, but is often studied in models of absorptive epithelia. Inherited disorders of amino acid transport are mostly severe and characterized by malabsorption syndromes and/or by aminoaciduria. These disorders often involve the failure of apical amino acid transport systems (e.g. Hartnup disorder) or, more rarely, the failure of basolateral transport systems, e.g. lysinuric acid intolerance [2].

The LAT2 (system L amino acid transporter 2) is composed of two polypeptides. The heavy-chain subunit CD98/4F2hc (Slc3a2) is a single-pass highly glycosylated protein, which associates through a disulfide bridge with a non-glycosylated 12-transmembrane-spanning light-chain subunit Lat2 (Slc7a8) [3]. The heavy chain is required for correct membrane localization of the complex [4]. Amino acid transport is mediated by the light-chain subunit, which is a member of the major facilitator superfamily. The complex is highly expressed along the renal proximal tubule epithelium and correlates with neutral amino acid reabsorption capacity along the nephron [5,6]. LAT2 bears broad substrate specificity, including all neutral amino acids except proline, and acts as an obligatory exchanger [5,7]. Antisense and overexpression studies in MDCK (Madin–Darby canine kidney) and in a hamster proximal tubule-derived cell line (OK cells) have indicated moderate changes of intracellular amino acid concentrations, and suggested a role in cysteine export [8,9]. The frog homologue of Lat1 was cloned as IU12, a developmentally regulated gene, mediating T3 (3,5,5′-tri-iodo-L-thyronine) and T4 (3,3′,5,5′-tetra-iodo-L-thyronine) uptake into Xenopus oocytes [10]. Human LAT2 (SLC7A8) and LAT1 (SLC7A5) also accept thyroid hormones in the Xenopus oocyte system [11]. Considering the large number of neutral amino acid transporters (Slc16a9, Slc6a15, Slc1a5, Slc6a14, Slc16a10, Slc43, Slc38, Slc7a10, Slc7a8, Slc7a5), it appears plausible that Lat2 and Lat1 have evolved to accommodate additional substrates, such as iodothyronines.

T3, as the most active form of iodothyronines, acts on nuclear receptors. Initially it was assumed that the hydrophobicity of thyroid hormones would allow passive diffusion through the lipid bilayer, despite the known zwitterionic character of the molecules [12]. The concept of facilitated diffusion of thyroid hormones was only widely adopted after the identification of patients with mutations in the most specific thyroid hormone transporter MCT8 (monocarboxylate transporter 8, SLC16A2) [1315]. Only then was it discovered that mutations in MCT8 are the cause of Allan–Herndon–Dudley syndrome (OMIM 300523), a severe X-linked mental retardation syndrome first described in 1944 [16]. Mice deficient in Slc16a2 replicate the endocrine phenotype of SLC16A2-deficient patients, but do not suffer from severe mental retardation. Since T3 uptake into primary neurons does not completely depend on Mct8, but can be competed with the Lat-specific substrate BCH [2-aminobicyclo-(2.2.1)-heptane-2-carboxylic acid], we have suggested that co-expression of Lat2 in neurons complements inactivation of Slc16a2 in mice [17]. Later, we could demonstrate that both Mct8 and Lat2 mediate thyroid hormone uptake into cultured astrocytes [18]. In the present study, we describe for the first time the impact of Slc7a8 deficiency on amino acid and thyroid hormone metabolism in mice.



Slc7a8-deficient mice on a C57Bl/6 genetic background were obtained from Jackson Laboratory. To produce the targeted allele, a β-galactosidase/neomycin phosphotransferase II-poly A-signal cassette (geo) was inserted into exon 2 of Slc7a8 replacing part of the coding sequence of transmembrane helices 1 and 2. The following primers were used for PCR genotyping: 5′-CAAATGCCAGCTGTCCTGACCTCAC-3′, 5′-CAGACTTAGGGATGGTGACGCCTAG-3′ and 5′-GGGTGGGATTAGATAAATGCCTGCTCT-3′. Mice were fed ad libitum with standard rodent breeding chow (Sniff) and were killed at approx. 12:00 h. Animals were kept in accordance with local regulations (Landesamt für Gesundheit und Soziales Berlin #T458/09), under standard conditions (12 h light/dark cycle) in a specific pathogen-free environment in the central animal facility of the Charité, Berlin. Analysed animals were always littermates from heterozygous intercrosses in order to avoid potential genetic drift between wild-type and mutant colonies.

Western blot analysis

Frozen mouse organs were powdered under liquid nitrogen, and cytosolic and membrane extracts for analysis were prepared as described previously [18]. Protein (80 μg) was separated by SDS/PAGE (12.5% gels). After transfer on to nitrocellulose membranes, membranes were incubated with primary antibodies [anti-Mct8, 1:500 dilution (ATLAS); anti-Lat2, 1:200 dilution (ImmunoGlobe); and anti-β-actin, 1:2000 dilution (Rockland)] overnight at 4°C, and afterwards with an HRP (horseradish peroxidase)-conjugated secondary antibody. Signals were detected using Amersham ECL (enhanced chemiluminescence) Plus (GE Healthcare) and Kodak Hyperfilm Plus as described previously [17].


Tissues were fixed in 4% (w/v) paraformaldehyde overnight and stained with Dako Cytomation Systems according to the manufacturer's instructions. Tissues were incubated with the primary anti-Lat2 antibody (1:250 dilution) overnight. Nuclei were counterstained with haematoxylin. Pictures were taken with a Zeiss Axioscope 2 MOT Plus equipped with an AxioCam MRc5 and Axiovision software. Alternatively, Cy2 (carbocyanine)- and Cy3 (indocarbocyanine)-labelled secondary antibodies (Jackson ImmunoResearch) were used to stain calbindin (1:1000 dilution; Swant) and parvalbumin (1:1000 dilution; Swant), and confocal images were taken with a Leica instrument at the Neuroscience Research Center core facility at Charité-Universitätsmedizin Berlin.

qPCR (quantitative PCR)

Total RNA of powdered mouse tissue was isolated using TRIzol® (Invitrogen). cDNA was synthesized using the iScript cDNA synthesis kit (Bio-Rad) according to the manufacturer's protocol. qPCR was performed using SYBR Green from Abgene (Thermo Scientific) on an iCycler (Bio-Rad). Primers used for qPCR detection of thyroid hormone transporters have been described previously [17].

Hormone measurements

TSH (thyroid-stimulating hormone), total serum T4 and total serum T3 were measured by immunoassays as described previously [19]. Per genotype, eight to twelve male animals were analysed.

Enzymatic activity measurements

Type I-deiodinase activity was determined as described previously [19]. Liver protein (10 μg) and kidney homogenate (15 μg) were used, as well as 0.1 μM unlabelled rT3 (reverse T3). Per genotype, three to four male animals were analysed. The measurement was repeated two to four times with similar results.

Amino acid analysis

Urine was collected in metabolic cages over a period of 24 h. Urinary and serum amino acids were determined in the Interdisciplinary Endocrinological Special Laboratory of Charité-Universitätsmedizin Berlin according to a routine HPLC method after derivatization and fluorescence detection. Norleucine was added as an internal standard. Although serum amino acid values agree well with reported data from other laboratories, absolute numbers of urinary amino acid concentrations may represent an overestimate, which may be caused by crowding of chromatograms. However, methods and results vary in the literature.


Targeted inactivation of Slc7a8 in mice

Lat2 is considered an important neutral amino acid transporter in kidney and intestine [2], but its physiological role in the intact animal is not known. Insertion of a geo cassette (see the Experimental section) into exon 2 of Slc7a8 disrupted the protein (Figure 1A). Homozygous mutant animals (Slc7a8−/−) were born at the expected Mendelian frequency and were identified by PCR genotyping (Figure 1B). Slc7a8−/− mice were indistinguishable from their wild-type or Slc7a8+/− littermates with respect to body weight and length (Figure 1C). Lat2 protein was expressed in kidney, brain, pituitary, testis, spleen (Figure 1D) and small intestine (Figure 1E). The Lat2 signal along the gastrointestinal tract was entirely abolished in Slc7a8−/− tissue (Figures 1E and 1F). Immunohistochemistry suggests that the faint signal observed in liver is derived from Kupffer cells (results not shown). The protein of 38 kDa detected in liver and brain is unrelated to Lat2 as it remains in Slc7a8−/− tissue. Expression of Lat2 in kidney is restricted to the basolateral membrane of proximal tubules (Figure 2A).

Figure 1 Gene targeting of Slc7a8 in mice

(A) Schematic representation of the murine Slc7a8 locus. Boxes designate exons. Exon 2 encoding parts of transmembrane helices 1 and 2 (white circles) was replaced with an in-frame geo cassette leading to a functional null allele. (B) PCR genotyping of mice using primers located as shown in (A). The black circle denotes the wild-type allele, the white circle denotes the targeted allele. (C) Comparison of body weight and body length between Slc7a8−/− and wild-type mice did not show any significant differences. n=8–12 male mice, age 60–250 days. (D) Expression of Lat2 in multiple tissues of wild-type mice. For Western blot analysis, 80 μg of protein from the membrane fraction was loaded per lane. β-Actin served as a control. Molecular mass markers are indicated on the right-hand side. (E) Expression of Lat2 along the gastrointestinal tract comparing wild-type and Slc7a8−/− mice. (F) Comparative Western blot analysis of Lat2 expression in liver, kidney and brain. β-Actin was used as a control. The specific band is labelled with an arrow. Mct8 protein expression is shown for comparison.

Figure 2 Amino acid analysis in Slc7a8−/− mice

(A) Immunohistochemical staining for Lat2 in kidney locates the protein in the basolateral membrane of proximal tubule epithelial cells. Counterstaining with haematoxylin. Scale bar is 100 μm. The signal is completely abrogated in Slc7a8−/− tissue. (B) Amino acid analysis in urine reveals neutral aminoaciduria in Slc7a8−/− mice. n=5–6 male mice per genotype. (C) Comparative qPCR analysis of amino acid and thyroid hormone transporters in the kidneys of wild-type and Slc7a8−/− mice. 18S rRNA was used for normalization. n=3–4 male mice per genotype. (D) Amino acid analysis in serum. n=4–7 male mice per genotype. Statistical analysis was calculated with ANOVA. *P< 0.05, **P< 0.01 and ***P< 0.001. n.d., not detected.

Aminoaciduria in Slc7a8−/− mice

Mutation of the related transporter SLC7A7 leads to lysinuric protein intolerance in patients [20] and a similar phenotype in Slc7a7−/− mice [21]. We therefore asked whether inactivation of Slc7a8 would similarly affect urinary amino acid excretion and reflect the substrate spectrum of Lat2 in vivo. Mice were placed in metabolic cages for 24 h and urine was collected. Urinary amino acid analysis revealed aminoaciduria in Slc7a8−/− mice (Figure 2B). Glycine, serine, threonine, glutamine, leucine and valine were increased in urine from mutant animals. Interestingly, some classical L-type amino acid transporter substrates such as isoleucine or aromatic amino acids were not affected. We therefore speculated that other amino acid transporters partially compensate for the loss of Lat2 transporters. Expression of several candidate transport proteins was compared by qPCR (Figure 2C). Although Slc7a8 mRNA was virtually undetectable, a slight increase in Slc16a10 (a transporter for aromatic amino acids) was observed, which was, however, only significant at the 5% level upon one-sided Student's t test. Amino acid profiles in serum were also altered in Slc7a8−/− animals. Several small neutral amino acids were increased in serum, partly resembling the pattern in urine. The most striking difference was an increase in lysine serum levels in the mutants (Figure 2D).

Serum thyroid hormone levels are unaltered in Slc7a8−/− mice

Lat2 is clearly capable of thyroid hormone transport in vivo [18] and is expressed in tissues responsive to thyroid hormones, including brain, pituitary and kidney. Inactivation of Lat2 may thus alter tissue responses to thyroid hormone or have an impact on circulating thyroid hormone levels. We therefore determined circulating levels of the thyroid hormones T3 and T4, as well as TSH, in a group of male animals (Figure 3). Neither of the hormones was altered in Slc7a8−/− mice.

Figure 3 Circulating thyroid hormone and thyrotropin levels remain normal in Slc7a8−/− mice

(Top panel) Total T3. (Middle panel) Total T4. (Bottom panel) TSH levels. Hormones were measured by immunoassays in sera from 8–12 male mice per genotype.

Thyroid-hormone-dependent gene activation

It is still possible that minor changes in thyroid hormone uptake into some tissues are not reflected by alterations in circulating hormone levels, but can be revealed on the level of target gene expression. We have therefore measured the activity of Dio1 (type I-deiodinase), a T3-responsive gene [22], in liver and kidney from Slc7a8−/− mice and wild-type littermates (Figures 4A and 4B). Again, no differences were observed. Impaired function of the thyroid axis is often reflected by changes in gene expression of Dio2 and Tshb in the pituitary. Again, no differences were observed by qPCR analysis, suggesting that the thyroid hormone axis functions normally in Slc7a8−/− mice.

Figure 4 Thyroid-hormone-responsive genes in liver, kidney and pituitary are unchanged in Slc7a8−/− mice

(A) Activity of hepatic Dio1 was determined in three to four male mice per genotype. No significant difference was observed (Student's t test). (B) Activity of renal Dio1 was determined in three to four male mice per genotype. No significant difference was observed (Student's t test). (C) Gene expression of Dio2 and Tshb in pituitary was assessed by qPCR in five male animals per genotype. 18S rRNA served as a reference gene. No differences were observed.

Brain development and impaired movement co-ordination

We then probed thyroid hormone signalling in the cerebral cortex and cerebellum by qPCR. RC3 (neurogranin), Hr (hairless) and Dio3 are known thyroid-hormone-responsive genes in the brain. None of these genes were differentially expressed in Slc7a8-mutant cerebellum (Figures 5A–5C) and cerebral cortex (see Supplementary Figure S1 at Gross histology did not reveal any apparent changes in brain size, structure and expression of neurochemical markers in somatosensory cortex and hippocampus (Supplementary Figure S2 at In particular, a reduction of parvalbumin expression in cortical interneurons is known to occur in mice which have experienced postnatal hypothyroidism [23] or are expressing a dominant-negative thyroid hormone receptor α1 [24]. Parvalbumin expression is unchanged in Slc7a8−/− cerebral cortex (Supplementary Figure S2). In contrast, movement co-ordination was slightly impaired in Slc7a8−/− mice, as revealed in the rotarod assay (Figure 5D). Cerebellar histogenesis was normal as judged by Nissl staining (Figures 5E and 5F) arguing against developmental hypothyroidism. In addition, immunohistochemical staining with markers for Purkinje and stellate cells was normal in Slc7a8 mutants (Figures 5G and 5H).

Figure 5 Cerebellar phenotype of Slc7a8−/− mice

Expression of genes responsive to thyroid hormone was assessed by qPCR in three to four cerebella per genotype. (A) RC3, neurogranin. (B) Hr, hairless. (C) Dio3. No significant differences were observed by Student's t test. (D) Rotarod analysis as a measure of movement co-ordination revealed a slight impairment in Slc7a8−/− mice as compared with controls. n=8–12 male mice per genotype. (E and F) Nissl staining of cerebellar sections does not reveal differences in histogenesis as usually observed in hypothyroid animals. (G and H) Immunohistochemical staining for calbindin (red), a Purkinje cell marker, and parvalbumin (green), a marker for Purkinje cells and cerebellar interneurons. Scale bars are 50 μm.


The phenotype of Slc7a8−/− mice is remarkably mild. In contrast, mutations of related transporters Slc7a9 and Slc7a7 in mice cause cystinuria [25] and lysinuric protein intolerance [21]. Based on the Slc7a8/Lat2 substrate spectrum, Bröer has proposed SLC7A8 as a candidate gene for isolated cystinuria (OMIM 238200) [2]. Although we have detected aminoaciduria in the mutants, their phenotype does not support this idea; in contrast, cysteine excretion is diminished in our mutants.

Lat2 is not exclusively, but most strongly, expressed in absorptive epithelia, kidney proximal tubules, small intestine and placenta. Accordingly, aminoaciduria is observed in Slc7a8−/− mice. Interestingly, only a subset of known Lat2 substrate amino acids is affected, i.e. excluding aromatic amino acids. This finding suggests that (i) aromatic amino acid transport is fully maintained by an aromatic amino acid transporter present in kidney tubular cells, and (ii) that remaining neutral amino acid transport is partially rescued by an alternative transporter. In the kidney, it has been proposed that Lat2 co-operates with two other transport activities expressed on the basolateral membrane [26]. A unidirectional exporter is required to drive amino acid export from the proximal tubule epithelium, because Lat2, as an obligate exchanger, cannot create a net flux over the membrane. The role of Lat2 is thus to extend the substrate spectrum for export, if the substrate spectrum of the uniporter is more limited [2,27]. A good candidate for such a uniporter is Tat1/Mct10/Slc16a10, an aromatic amino acid transporter. Compatible with this idea, we have found moderate induction of Slc16a10 mRNA in kidneys of Slc7a8-mutant animals. Influx of exported aromatic amino acids may then allow the export of the whole spectrum of Lat2 substrate amino acids. Some of these amino acids are then used in turn as co-substrates for Slc7a7, the y+Lat1 transporter of basic amino acids mutated in lysinuric protein intolerance. Functional interference of Slc7a8 with basic amino acid transport is supported by increased lysine serum levels in Slc7a8 mutants. However, the mechanism cannot be inferred from the data presently available. Intestinal amino acid absorption has not been tested in these mice, since it is apparently not limiting considering normal growth and weight gain of mutants. Moreover, Lat2 has been implicated in amino acid transport across the placenta [28]. Normal frequency of Slc7a8 mutants among offspring and normal growth together argue in favour of alternative placental transporters, although we have not formally investigated this question by, for example, weighing Slc7a8−/− mouse embryos or newborn offspring from Slc7a8−/− mothers. This finding is in sharp contrast with the fetal and growth phenotype of Slc7a7−/− mice [21]. It appears likely that other neutral amino acid transporters can compensate for many Slc7a8 functions in vivo. Limiting the dietary availability of essential Lat2 substrate amino acids may help reveal a phenotype compensated in our experiments by affluent amounts of amino acids in our breeding diet.

The high abundance of Lat2 in the brain suggests some role for the protein in the brain. This notion is supported by the moderate impairment of movement co-ordination in Slc7a8 mutants revealed in the rotarod task. Bröer et al. [29] proposed a possible role for Lat2 in alanine uptake into astrocytes. Alanine was suggested as the source of ammonia in the cerebral glutamine/glutamate cycle [29], which also involves a net transfer of ammonia to neurons. In this cycle, astrocytes take up extracellular glutamate released by excitatory neurons and convert it into glutamine, which is then provided back again to neurons as a source of glutamate [30]. Inactivation of Lat2 may thus slow down the glutamine/glutamate cycle by reducing the flux of alanine. Alternatively, or in addition, Lat2 may participate in astroglial glutamine export or neuronal glutamine import, although it is currently assumed that these activities are covered by systems A and N [31]. The exceptionally high concentration of glutamine in cerebrospinal fluid would make Lat2 a good candidate for glutamine uptake into neurons. It is beyond the scope of the present study to distinguish the relative roles of systems L, N and A in the glutamine/glutamate cycle. A dedicated neurochemical study would be needed to address this issue.

Lat2 accepts thyroid hormones in vitro [10,11,32] and in vivo [18]. T3 and T4 uptake into primary Slc7a8−/− astrocytes is significantly reduced, to a level identical with uptake in the presence of BCH, a specific Lat inhibitor [18]. We have analysed thyroid hormones and thyroid-hormone-dependent gene expression in organs from Slc7a8−/− mice, because we had proposed a possible compensation of thyroid hormone transport by Lat2 in Slc16a2-deficient mice [17]. Plasma T3 levels are increased and T4 levels are decreased in Slc16a2−/y mice [33,34]. Both Mct8 and Lat2 are expressed in kidney, brain and pituitary. Our findings of normal T3, T4 and TSH levels in Slc7a8−/− mice suggest that cells determining these hormone levels depend on Mct8, but not Lat2. Thus a feature of these critical cells may be lack of co-expression of Mct8 and Lat2. Since urinary loss of thyroid hormones in Slc16a2-deficient mice may reduce T4 levels [35], we have attempted to determine urinary thyroid hormone loss in Slc7a8−/− mice by a liquid chromatography-tandem MS method [36]. Although T4 was clearly detectable, the signal was close to background and thus could not be reliably quantified. Normal activity of renal Dio1 is in contrast with elevated activity in Slc16a2−/y kidney. This finding suggests that intracellular T3 signalling is not changed in Slc7a8−/− kidney.

Normal thyroid hormone levels in the mutants are in agreement with normal (thyroid-hormone-dependent) cerebral and cerebellar development in Slc7a8 mutants. Moreover, T3-dependent gene expression is not changed. We conclude that there is no indication at the moment for insufficient thyroid hormone transport into brain cells. This could be a consequence of co-expression in neurons and astrocytes of Mct8 and Lat2 [17,18] and would be consistent with a lack of overt neurological phenotype in Slc16a2-deficient mice, which contrasts with the devastating neurological disorder in patients lacking functional MCT8. It will be interesting to test, by compound inactivation of both genes, the hypothesis that Lat2 and Mct8 compensate for each other in thyroid hormone transport into brain cells.


Doreen Braun, Eva Wirth, Franziska Wohlgemuth, Nathalie Reix and Ulrich Schweizer obtained the data; Doreen Braun, Eva Wirth, Nathalie Reix, Marc Klein, Josef Köhrle and Ulrich Schweizer analysed and interpreted the data; Annette Grüters, Josef Köhrle and Ulrich Schweizer designed the experiments; Doreen Braun, Josef Köhrle and Ulrich Schweizer drafted the paper; all authors contributed to the writing of and approved the final version of the paper.


This work was supported by the Deutsche Forschungsgemeinschaft [project numbers SFB665 TP A7-2, GK1208] and Charité-Universitätsmedizin Berlin.

Abbreviations: BCH, 2-aminobicyclo-(2.2.1)-heptane-2-carboxylic acid; Dio, deiodinase; LAT2/Lat2, system L amino acid transporter 2; MCT8/Mct8, monocarboxylate transporter 8; qPCR, quantitative PCR; T3, 3,5,5′-tri-iodo-L-thyronine; T4, 3,3′,5,5′-tetra-iodo-L-thyronine; TSH, thyroid-stimulating hormone


View Abstract