SGLT5 is the renal transporter for 1,5-anhydroglucitol, a major player in two rare forms of neutropenia

Abstract

Neutropenia and neutrophil dysfunction in glycogen storage disease type 1b (GSD1b) and severe congenital neutropenia type 4 (SCN4), associated with deficiencies of the glucose-6-phosphate transporter (G6PT/SLC37A4) and the phosphatase G6PC3, respectively, are the result of the accumulation of 1,5-anhydroglucitol-6-phosphate in neutrophils. This is an inhibitor of hexokinase made from 1,5-anhydroglucitol (1,5-AG), an abundant polyol in blood. 1,5-AG is presumed to be reabsorbed in the kidney by a sodium-dependent-transporter of uncertain identity, possibly SGLT4/SLC5A9 or SGLT5/SLC5A10. Lowering blood 1,5-AG with an SGLT2-inhibitor greatly improved neutrophil counts and function in G6PC3-deficient and GSD1b patients. Yet, this effect is most likely mediated indirectly, through the inhibition of the renal 1,5-AG transporter by glucose, when its concentration rises in the renal tubule following inhibition of SGLT2. To identify the 1,5-AG transporter, both human and mouse SGLT4 and SGLT5 were expressed in HEK293T cells and transport measurements were performed with radiolabelled compounds. We found that SGLT5 is a better carrier for 1,5-AG than for mannose, while the opposite is true for human SGLT4. Heterozygous variants in SGLT5, associated with a low level of blood 1,5-AG in humans cause a 50-100% reduction in 1,5-AG transport activity tested in model cell lines, indicating that SGLT5 is the predominant kidney 1,5-AG transporter. These and other findings led to the conclusion that (1) SGLT5 is the main renal transporter of 1,5-AG; (2) frequent heterozygous mutations (allelic frequency > 1%) in SGLT5 lower blood 1,5-AG, favourably influencing neutropenia in G6PC3 or G6PT deficiency; (3) the effect of SGLT2-inhibitors on blood 1,5-AG level is largely indirect; (4) specific SGLT5-inhibitors would be more efficient to treat these neutropenias than SGLT2-inhibitors.

Keywords: 1,5-Anhydroglucitol; Empagliflozin; G6PC3-deficiency; GSD1b; Glycogen storage disease type Ib; Neutropenia; SCN4; SGLT2-inhibitors; SGLT4; SGLT5.

Fig. 1

Transport of 1,5-AG and mannose by HEK293T cell overexpressing human and mouse SGLT4 and SGLT5 isoforms. A Alignment of human SGLT5/SLC5A10 protein sequences highlighting the 16 amino acid non-conserved N-terminal extension of exon 10 present in SGLT5-iso1. The sequences blasted correspond to: human SGLT5-iso2 (NP_001035915.1) and SGLT5-iso1 (NP_689564.3); mouse SGLT5 (NP_001028399.1); zebrafish SGLT5. B Impact of the 6xHis-tag on the transport activity of cells transfected to transiently overexpress SGLT4 or SGLT5. Transport was measured in 24 well plates with 0.6 × 10⁶ cells during 30 min at 37 °C in 5% CO₂ in the presence of 10 µM 2-[¹H + ³H]-1,5-AG (upper panel) or U-[¹²C + ¹⁴C]-mannose (lower panel) as described in Materials and Methods. C Transport activity for 1,5-AG (upper panel) and mannose (lower panel) of model cell lines overexpressing the untagged transporters in a stable fashion created by lentiviral transduction of HEK293T cells as described in Materials and Methods. Transport activities were measured as in (B) for 60 min. Data corresponds to n = 3 of least two independent experiments. 1,5-AG–1,5-anhydroglucitol

Fig. 2

Human SGLT5 transports 1,5-AG while human SGLT4 transports mannose and not 1,5-AG. Saturation curves for the transport of A 1,5-anhydroglucitol and mannose by SGLT5 and SGLT4. Transport was measured during 30 and/or 60 min at 37 °C in 5% CO₂ in 24 well plates with 0.6 × 10⁶ cells overexpressing human SGLT4 or SGLT5 in the presence of shown concentrations of 2-[¹H + ³H]-1,5-AG (left panel) and U-[¹²C + ¹⁴C]-mannose (right panel). B Kinetic constants for both transporters were computed by fitting the data to the Michaelis–Menten model in Prism – GraphPad. The catalytic efficiency is estimated by the Vmax/Km values. Data corresponds to n ≥ 3 in at least 3 independent experiments. 1,5-AG – 1,5-anhydroglucitol

Fig. 3

Substrate specificity of human SGLT4 and SGLT5. Inhibition curves of the transport activity of A 1,5-anhydroglucitol by human SGLT5 and B mannose by human SGLT4. Measurements were performed during 30 and/or 60 min in the presence of 10 µM 2-[¹H + ³H]-1,5-AG or U-[¹²C + ¹⁴C]-mannose and the indicated concentrations of the inhibitor sugars in 24 well plates with 0.6 × 10⁶ cells overexpressing human SGLT5 or SGLT4, respectively. C When shown, the IC₅₀ values are derived in Prism – GraphPad from the curves in A and B and indicate the concentration of the sugars needed to inhibit by 50% the transport activity. When the inhibition was too weak and IC₅₀ values could not be derived, the degree of inhibition with 10 mM inhibitor is shown. The structures of the various inhibitory sugars used are shown (Haworth projections) to highlight the absence of the OH group bound to carbon 1 (C1) in 1,5-anhydromannitol, 1,5-anhydroglucitol and 1,5-anhydrofructose, and the orientation of the OH group bound to carbon 2 (left OH shown in red; right OH shown in green). The structure of the molecule of fructose is shown after a rotation along the oxygen 6–carbon 4 axis to indicate its structure similarity with 1,5-anhydromannitol. Data corresponds to n = 3 in at least 2 independent experiments. 1,5-AM – 1,5-anhydromannitol; 1,5-AG–1,5-anhydroglucitol; 1,5-AF – 1,5-anhydrofructose

Fig. 4

Site directed mutagenesis of specific features of SGLT5 binding pocket. A Multiple sequence alignment of selected human SGLT5 homologs showing strictly conserved residues in the binding pocket that possibly interact with the OH groups carried by carbons 1 and 2 of the hexose substrate molecules. The indicated residues in SGLT5 (1-Ser70, 2-Glu71 and 3-Leu75) were replaced individually or all together (as shown in panel C) by the corresponding residues in SGLT1 and SGLT2 (1-Asn78/75, 2-Ile79/76 and 3-His83/80). The protein sequence of the homologous proteins were aligned using Clustal Omega (https://www.ebi.ac.uk/Tools/msa/clustalo/). B SGLT2-MAP17 cryo-EM structure (code PDB: 7vsi [33]) of the substrate binding pocket crystalized with empagliflozin in it (for clarity we hid the aglycone part of the empagliflozin molecule which is normally linked to carbon 1), showing the position of 1-Asn75, 2-Ile 76 and 3-His80 and depicting their possible interactions (black lines: interactions shown in the structure proposed by Niu et al. [33]; red lines: interactions suggested in Sala-Rabanal et al. [30]) with the substrate and other residues around the binding pocket. C Impact of replacing the indicated residues in SGLT5 by the equivalent ones in SGLT1 and SGLT2 on the transport activities of 1,5-AG and mannose. SGLT5-TM corresponds to a mutant SGLT5 carrying all three mutations (S70N–E71I–L75H). Transport activities were measured in the presence of 50 µM 2-[¹H + ³H]-1,5-AG or U-[¹²C + ¹⁴C]-mannose. Data corresponds to n = 3 in at least 3 independent experiments. 1,5-AG–1,5-anhydroglucitol

Fig. 5

Impact of mutations in SGLT5 on the transport activity for 1,5-AG and on the affinity for gliflozins. A Saturation curves for the transport of 1,5-AG by various model cell lines overexpressing the indicated SGLT5 mutants (see Table 1). 1,5-AG transport was measured in the presence of the indicated concentrations of 2-[¹H + ³H]-1,5-AG during 30 and/or 60 min. Data corresponds to n = 3 in at least 2 independent experiments. B Localisation and illustration of the possible impact of the mutations present in the variants of human SGLT5-iso2 (NP_001035915.1) modelled using AlphaFold [34, 35]. C Inhibition curves of the transport activity of 1,5-AG by human SGLT5 measured in the presence of 10 µM 2-[¹H + ³H]-1,5-AG and increasing concentrations of the indicated gliflozins. The IC₅₀ values are derived in Prism – GraphPad from the inhibition plots shown and indicate the concentration needed of each gliflozin to inhibit by 50% the transport activity. D Impact of mutations found in SGLT5 variants on the ability of empagliflozin to inhibit the transport of 1,5-AG. Data corresponds to n = 3 in at least 2 independent experiments. EMPA - empagliflozin, DAPA - dapagliflozin, REMO - remogliflozin and 1,5-AG–1,5-anhydroglucitol