Deficiency of the sedoheptulose kinase (Shpk) does not alter the ability of hematopoietic stem cells to rescue cystinosis in the mouse model

Abstract

Cystinosis is an autosomal recessive lysosomal storage disorder caused by mutations in the CTNS gene encoding the lysosomal cystine transporter, cystinosin, and leading to multi-organ degeneration including kidney failure. A clinical trial for cystinosis is ongoing to test the safety and efficacy of transplantation of autologous hematopoietic stem and progenitor cells (HSPCs) ex vivo gene-modified to introduce functional CTNS cDNA. Preclinical studies in Ctns^−/− mice previously showed that a single HSPC transplantation led to significant tissue cystine decrease and long-term tissue preservation. The main mechanism of action involves the differentiation of the transplanted HSPCs into macrophages within tissues and transfer of cystinosin-bearing lysosomes to the diseased cells via tunneling nanotubes. However, a major concern was that the most common cystinosis-causing mutation in humans is a 57-kb deletion that eliminates not only CTNS but also the adjacent sedopheptulose kinase SHPK/ CARKL gene encoding a metabolic enzyme that influences macrophage polarization. Here, we investigated if absence of Shpk could negatively impact the efficiency of transplanted HSPCs to differentiate into macrophages within tissues and then to prevent cystinosis rescue. We generated Shpk knockout mouse models and detected a phenotype consisting of perturbations in the pentose phosphate pathway (PPP), the metabolic shunt regulated by SHPK. Shpk^−/− mice also recapitulated the urinary excretion of sedoheptulose and erythritol found in cystinosis patients homozygous for the 57-kb deletion. Transplantation of Shpk^−/−-HSPCs into Ctns^−/− mice resulted in significant reduction in tissue cystine load and restoration of Ctns expression, as well as improved kidney architecture comparable to WT-HSPC recipients. Altogether, these data demonstrate that absence of SHPK does not alter the ability of HSPCs to rescue cystinosis, and then patients homozygous for the 57-kb deletion should benefit from ex vivo gene therapy and can be enrolled in the ongoing clinical trial. However, because of the limits inherent to animal models, outcomes of this patient population will be carefully compared to the other enrolled subjects.

1. Introduction

Cystinosis is an autosomal recessive lysosomal storage disorder caused by mutations in the CTNS gene that affects one out of every 100,000 to 200,000 live births [1]. CTNS encodes a proton-driven transmembrane lysosomal cystine transporter, cystinosin [2]. Over 100 unique mutations have been identified that perturb the localization or function of the cystinosin protein, ranging from single base substitutions to large indels [2–4]. Loss of the ubiquitously expressed cystinosin leads to lysosomal cystine accumulation and crystallization throughout the body, causing multisystemic organ failure [5]. Renal Fanconi syndrome presents in infants as the first clinical manifestation, and patients progressively develop end-stage renal failure, photophobia, myopathy, diabetes, hypothyroidism and neurological defects [1,6]. The main targeted therapeutic is the drug cysteamine, which facilitates lysosomal cystine export, prolonging life expectancy to early adulthood [7]. However, severe side-effects and extremely frequent and burdensome dosing lead to issues with patient compliance and, ultimately, this treatment merely delays symptomatic progression [8].

We pioneered a new treatment to address the root genetic cause of cystinosis by utilizing a hematopoietic stem and progenitor cell (HSPC) gene therapy approach [9]. Extensive preclinical experimentation using the Ctns^−/− mouse model established that a single transplantation of wildtype (WT) HSPCs lead to lifelong morphological and functional preservation of affected tissues including the kidney, cornea and thyroid [10–13]. Mechanism of action studies revealed that therapeutic efficacy relies upon HSPC differentiation into macrophages within tissues and delivery of cystinosin-bearing lysosomes to the diseased cells via open-ended membranous protrusions known as tunneling nanotubes (TNTs) [14,15]. Promising results in mice led to the initiation of a phase 1/2 clinical trial in which cystinosis patients undergo autologous transplantation of HSPCs, modified ex vivo using a lentiviral vector to express functional CTNS (ClinicalTrials.gov Identifier #NCT03897361).

However, the mechanism of action underlying HSPC-mediated tissue preservation in cystinosis raised a concern for subject selection in the clinical trial. Indeed, the most common mutation in cystinosis is a large 57 k base pair (kb) deletion spanning not only the CTNS gene, but also the entirety of the adjacent sedoheptulokinase (SHPK, a.k.a CARKL) locus [16]. About 40% of human cystinosis patients are homozygous for the 57-kb deletion and so lack any functional copy of SHPK, a metabolic enzyme that regulates the formation of sedoheptulose-7-phosphate (S7P) which is a key component of the non-oxidative arm of the pentose phosphate pathway (PPP) [17]. Alterations of PPP products and intermediates have been detected in patients’ urine and blood spots [18,19]. Importantly, SHPK was shown to regulate macrophage polarization and differentiation by altering the PPP flux [20]. As HSPC therapy relies on differentiation of functional macrophages, it is critical to determine if patients carrying the 57-kb deletion at the homozygous state would still benefit from the stem cell gene therapy approach.

To address this issue, we generated and characterized two novel Shpk-deficient mouse models (Shpk^−/− mice). These mice exhibit alterations of PPP in liver and urine. We transplanted Ctns^−/− mice with HSPCs isolated from Shpk^−/− mice and demonstrated their ability of rescue cystinosis as assessed by reduction in tissue cystine accumulation, restoration of Ctns expression, and preservation of renal morphology. Our results indicate that cystinosis patients lacking SHPK are equally likely to receive therapeutic benefit from autologous transplantation of gene-modified HSPCs and will be enrolled in the clinical trial.

2. Material and methods

2.1. Mice and ethics statement

C57BL/6 J Ctns^−/− mice were provided by Dr. Antignac (Inserm U983, Paris, France). Shpk^ΔATG (C57BL/6 J Shpk^−/−) mice were generated at the UCSD Mouse Genomics Core using CRISPR/Cas9 and sgRNA targeting the start codon of Shpk genomic locus. Shpk^ΔE2 (C57BL/6 Shpk^−/−) mice were generated at the Mutant Mouse Resource & Research Center at UC Davis using CRISPR/Cas9 and sgRNA targeting exon 2 of Shpk. Both strains were also crossbred with transgenic DsRed (B6.Cg-Tg(CAG-DsRed*MST)1Nagy/J) mice provided by the Jackson Laboratory (Bar Harbor, ME) to establish colored lines with constitutive expression of DsRed fluorescent protein (DsRed⁺Shpk^−/− mice). All strains and mouse procedures were approved by the University of California, San Diego (UCSD) in accordance with the guidelines set forth by the Institutional Animal Care and Use Committee (Protocol ID #S12288).

2.2. Mouse genotyping

Shpk mutant status was identified using PCR-based genotyping assays from mouse tail tissue. DirectPCR (Tail) (Viagen) and Protinease K treatment of mouse tails was used for DNA extraction. PCR protocol of Shpk^ΔATG DNA amplification is as follows: 95 °C (3 min) initial denaturation; 30 cycles of 98 °C (20 s) denaturation, 65 °C (15 s) annealing, and 72 °C (15 s) extension; and 72 °C (1 min/kb) final extension. Resultant wild-type band sizes were 632 bp; and mutant size was 495 bp. PCR protocol of Shpk^ΔE2 DNA amplification was provided by UC Davis and is as follows: 94 °C (2 min) initial denaturation; 10 cycles of 94 °C (10 s) denaturation, 65 °C (15 s) annealing with consecutive runs lowering in temperature by 1 °C (−1 °C/cycle), and 68 °C (2 min) extension; and 25 cycles of 94 °C (2 min) denaturation, 65 °C annealing (30 s), and 68 °C (2 min) extension with consecutive runs increasing in duration by 20 s (+20 s/cycle). Resultant wild-type band sizes were 638 and 1184 bp; and mutant size was 500 bp. Primer sequences can be found in Table S1.

2.3. Real time-qPCR

RNeasy Mini Kit and protocol (Qiagen, Hilden, Germany) was used for whole RNA extraction; blood RNA was extracted using the RNeasy Protect Animal Blood Kit (Qiagen). cDNA synthesis occurred by reverse transcription of 1 μg tissue and blood RNA using iScript cDNA Synthesis Kit (Bio-Rad, Hercules, CA). RT-qPCR reaction assembly consisted of 5 ul iTaq Universal SYBR Green, 3 μL cDNA (5 ng/ul), 0.6 μL forward and reverse primer (10 μM), and 1.8 μL H20, and ran on a CFX96 thermocycler (Bio-Rad) under the following conditions: 95 °C (30 s); 40 cycles of 95 °C (5 s) and 60 °C (30 s); then 65 °C (5 s); and 95 °C (5 s). Gene expression was measured using the delta/delta CT method relative to glyceraldehyde 3-phosphate dehydrogenase (Gapdh). All primer sequences are shown in Table S1.

2.4. Immunoblotting

RIPA lysis buffer (Sigma, St Louis, MO) composed of protease/phosphatase inhibitor cocktail (Thermo Fisher Scientific, Waltham, MA) was used for protein lysis. The Pierce BCA Protein Assay Kit (Thermo Fisher Scientific) was used to calibrate protein concentrations. SDS-NuPAGE treatment of 10–20 μg of protein were ran on 4–12% Bis-Tris Gel (Novex by Life Technologies, Carlsbad, CA) and transferred to PVDF membranes. Primary incubation with anti-Shpk antibody (Sigma; 1:2000) ran at 4 °C overnight. The next day, membranes were washed three times in Protein Wash Buffer (Azure Biosystems, Dublin, CA) and incubated in HRP-conjugated secondary antibodies goat anti-rabbit and rhodamine anti-tubulin (BioRad; 1:2000) for 1 h at RT. Membranes are again washed three times and a final time in 1× tris buffered saline (TBS). Blots were activated by HRP substrate Radiance Plus ECL Reagent (Azure) and chemiluminescent detection was captured on the c600 Imager (Azure). Western blots were quantitated by comparative densitometry against tubulin loading control.

2.5. Metabolomics

Harvested livers were rapidly freeze-clamped and placed in ice-cold liquid nitrogen to preserve robust metabolic activity. Samples were then powderized and 20 mg was resuspended in extraction solution (40% methanol/40% acetonitrile/20% H₂0) supplemented with 40 μL ^13C-labeled internal standards (Cambridge Isotype Laboratories) per mL solution. Samples were homogenized using Precellys® homogenizer, immediately returned to dry ice for 1 h, then spun at 13000 xg for 15 mins. Supernatant was then lyophilized on a speed vac with cold trap to dryness, and then injected onto an LC-MS/MS system (API 4500 triple quadrupole MS, AB Sciex) for quantification of metabolites of interest (PPP intermediates, glycolytic intermediates, and nucleotides), as thoroughly described in the supplementary methods. Multiple batches from different collections were scaled together using WT controls, with a scale factor applied that preserved the intra-batch fold changes between the different cell types. This data is presented in the box plots in terms of relative abundance in relation to the controls. Urine analysis of erythritol, erythronic acid, and sedoheptulose was also performed using a separate LC-MS/MS method, as also described in the Supplementary Methods. Calibration curves, chromatograms, and QC performance for this method are found in Fig. S3.

2.6. HSPC isolation and transplantation

HSPC isolation and transplantation were performed as previously described [9]. In brief, bone marrow fluid was flushed from harvested tibias and femurs of donor wild-type, Ctns^−/−, Shpk^ΔATG, and Shpk^ΔE2 mice. Sca1⁺ HSPCs were isolated by immunomagnetic separation using magnetic beads conjugated to anti-Sca1 antibody (Miltenyi Biotec, Auburn, CA). Ctns^−/− mice were lethally irradiated (6.5 Gy; X-Rad 320, PXi) 24 h prior injection of 100 μL of PBS containing 1× 10⁶ HSPCs via tail vein. At ~6 months post-transplantation, recipient mice were sacrificed and kidney, liver, and spleen samples were isolated for analyses of therapeutic efficacy.

2.7. Cystine content

Tissue cystine was measured as previously described [9]. Briefly, explanted tissues were homogenized using the Precellys 24 homogenizer (Bertin Technologies) N-ethylmaleimide (Fluka Biochemika). Lysates were transferred to 15% % 5-sulfosalicylic acid (SSA) (Fluka Biochemika) and centrifuged to remove proteins for comparative quantitation using BCA assay (Pierce). The cystine-containing supernatants were transferred to 3% SSA and diluted before being quantitated by mass spectrometry at the UCSD Biochemical Genetics as described previously [21].

2.8. Histology

To assess general tissue morphology, dissected tissues were fixed in 10% formalin for 48 h and then transferred to the UCSD Moores Cancer Center Biorepository & Tissue Technology Center. Samples were embedded in paraffin, sectioned at 4 μm and stained with hematoxylin&Eosin (H&E) staining to assess histologic anomalies. For kidney, further analysis was done to quantify the extent of cortical damage as previously described [13]. To estimate the extend of the affected parenchyma in kidneys, periodic acidshift (PAS) staining was also performed, which stains basement membrane, so is useful for evaluating glomerular cell number, basement membrane, mesangium, tubular basement membrane and apical end of the proximal convoluted tubule epithelium’s brush border [22]. All kidney’s H&E and PAS slides were blindly scored by Dr. Valeria Estrada (Tissue Technology Shared Resource, Moores Cancer Center, UCSD) based on extent of cortical damage; a grade of 1 corresponded to 0–10% damage, Grade 2 to 10–30%, Grade 3 to 30–50% and Grade 4 to 50–70% [13].

2.9. Immunofluorescence

Tissues were fixed in 4% paraformaldehyde, dehydrated in 20% sucrose and mounted in Tissue Tek Optimum Cutting Temperature (OCT) stored at −80 °C. Tissues were cut into 16 μm-wide sections and placed on IHC slides. For staining, slides were washed in PBS three times for 10 min at RT, treated with blocking buffer (1% BSA, 5% goat serum, 0.3% triton X-100) for 1 h at RT, and probed with macrophage markers anti-F4/80 (Santa Cruz; 1:50) and anti-MHCII (BD Biosciences; 1:50) primary antibodies overnight at 4 °C. The following day, slides were washed in PBS three times and probed with secondary antibodies anti-rabbit AF562 and phalloidin AF647 for 1 h at RT. After two more 10-min PBS washes and a final PBS and DAPI (1:1000) wash, slides were mounted with glass slips using Prolong Gold. The following day, images were taken using the LSM 880 confocal microscope (Zeiss).

2.10. Renal function

24-h urine collections were preformed using metabolic cages with collection over ice and treated with 100× Protease Inhibitors. Serum was obtained from survival retro-orbital bleeding and placed in heparin tubes for blood chemistry including creatinine, albumin, urea, phosphorus and protein, and in EDTA tubes for complete blood count; both being measured at UCSD ACP Veterinary Diagnostic Laboratory Service. Urine creatinine, phosphate and glucose were measured using colorimetric assays according to the manufacturer recommendations (BioAssay Systems, Hayward, CA). Total proteinuria was measured using Pierce BCA Protein Assay Kit (Thermo Fisher Scientific).

2.11. Statistics

All statistics were analyzed using Graphpad Prism 9 software. Briefly, Student’s t-test was used for comparisons of two groups, and one-way analysis of variance (ANOVA) for comparisons of three or more conditions with Tukey’s multiple comparison test used to evaluate differences between individual groups. All graphs display mean ± standard error of the mean (sem).

3. Results

3.1. Shpk knockout generation and expression characterization

To obtain a source of Shpk-deficient HSPCs, we generated two novel Shpk knockout (Shpk^−/−) mouse models using CRISPR-Cas9 guides to create heritable homozygous indels via non-homologous end joining (Fig. 1A). We observed stable and heritable transmission of two Shpk genomic frameshift deletions; a 168-bp deletion centered around the start codon (Shpk^ΔATG), and a 675-bp elimination of the entire second exon (Shpk ^ΔE2) (Fig. S1 and Table 1 for primer sequences). Heterozygote mice were bred and Shpk^−/− mice were obtained following a mendelian ratio in both models.

Fig. 1.

Generation and characterization of two novel Shpk knockout mouse models. (A) Diagram of the mouse genomic region between Ctns and Shpk. CRISPR-Cas9 guides depicted at start codon (Shpk^ΔATG) or second exon (Shpk^ΔE2). CDS: coding DNA sequence, TSS: transcription start site. (B) Shpk expression as measured by qPCR in various tissues from the Shpk-knockout mice as compared to WT and Ctns^−/− mice. Data are mean ± standard deviation (SD) (C-D) Representative western blot showing SHPK expression in liver (C) and kidney (D) from WT, Ctns^−/−, Shpk^ΔATG, and Shpk^ΔE2 mice with anti-tubulin as loading control. Quantitation is based on relative densitometry between size-appropriate SHPK band and tubulin. Data are mean ± sem. * P < 0.05, ** P < 0.01, *** P < 0.001 relative to WT.

Table 1.

Renal function markers in the urine and serum.

	WT (n = 3)	Ctns−/− (n = 5)	Ctns−/− HSPC (n = 8)	WT HSPC (n = 11)	Shpk−/− HSPC (n = 18)	Reference Ranges*
Serum
Creatinine (mg/dL)	0.2 ± 0.0	0.2 ± 0.0	0.2 ± 0.0	0.3 ± 0.0	0.2 ± 0.0	0.2–0.8
Blood Urea Nitrogen (mg/dL)	17.7 ± 3.7	23 ± 2.0	29.4 ± 1.9	31.7 ± 4.4	30.1 ± 3.4	18–29
Phosphorus (mg/dL)	5.9 ± 1.3	4.7 ± 0.5	5.7 ± 0.3	5.6 ± 0.3	6.6 ± 0.4	6.1–10.1
Albumin (g/dL)	3.7 ± 0.2	3.6 ± 0.1	3.5 ± 0.1	3.5 ± 0.1	3.6 ± 0.1	2.5–4.8
Total Protein (g/dL)	4.4 ± 0.3	4.4 ± 0.2	4.7 ± 0.2	4.6 ± 0.1	4.7 ± 0.1	3.6–6.6
	WT (n = 4)	Ctns−/− (n = 7)	Ctns−/− HSPC (n = 12)	WT HSPC (n = 13)	Shpk−/− HSPC (n = 27)	Reference Ranges*
Urine
Polyuria (mL)	1.04 ± 0.19	3.31 ± 0.79a	1.20 ± 0.26b	1.41 ± 0.25b	1.70 ± 0.28b
Phosphate/creatinine (μmol/μmol Cr)	25.53 ± 4.50	31.61 ± 4.51	31.30 ± 5.55	31.43 ± 3.63	39.35 ± 6.74
Protein/creatinine (mg/mgCr)	0.60 ± 0.03	3.66 ± 1.89	6.35 ± 1.91	4.42 ± 1.86	3.97 ± 1.03
Glucose/creatinine (mg/mgCr)	2.56 ± 0.93	17.48 ± 5.55	13.78 ± 3.34	6.26 ± 0.97	10.43 ± 3.62

Data are presented as mean ± sem.

Urine data are from 24-h urine measures and reported to urine creatine ratio.

^*Normal range for C57BL/6 mice – Charles River.

^ap < 0.05 vs. WT mice.

^bp < 0.05 vs. Ctns^−/− mice.

We then characterized Shpk mRNA and protein expression as a measure of knockout efficacy. Real-time quantitative-PCR (RT-qPCR) analysis confirmed a dramatic reduction of Shpk expression in Shpk^ΔE2 tissues compared to WT controls (Fig. 1B). In contrast, Shpk mRNA expression in Shpk^ΔATG tissues retained variable levels of transcription with no significant differences in any tissue compared to WT samples. However, by immunoblotting we detected a strong size-appropriate band of 57-kilodaltons (kD) in WT and Ctns^−/− liver and kidney that were completely absent in both Shpk^ΔE2 and Shpk^ΔATG mice (Fig. 1C–D). These results indicate that despite residual mRNA expression in Shpk^ΔATG, neither knockout mouse model resulted in any translated SHPK protein. Surprisingly, significantly decreased SHPK expression was detected in the Ctns^−/− liver and kidney compared to WT (Fig. 1D), the reason of these results is unknown. SHPK protein was not detected in WT brain, heart, spleen, and muscle (data not shown). Histological analysis of kidney and liver from Shpk^ΔE2 and Shpk^ΔATG mice did not reveal any obvious difference compared to WT mice (data not shown). Note that this model does not recapitulate the patients’ 57 kb-deletion that remove both CTNS and SHPK genes. However, the because the gene therapy approach restores CTNS in patients’ HSPCs, they also become only SHPK-deficient.

3.2. Pentose Phosphate Pathway (PPP) metabolomics are disrupted in Shpk^−/− liver and urine

Phosphorylation of sedoheptulose into sedoheptulose-7-phosphate (S7P) via SHPK enzymatic activity is an important step in the pentose phosphate pathway (Fig. 2A) [23,24]. We assessed PPP metabolite abundance by intracellular mass spectrometry in flash-frozen liver samples to determine if genetic ablation of Shpk resulted in a metabolic phenotype. We specifically focused on the following PPP metabolites: i) S7P as the direct Shpk product; ii) glyceraldehyde-3-phosphate as an aldose intermediate (measured together with its isomer dihydroxacetone phosphate); iii) erythrose-4-phosphate (E4P), a downstream erythritol precursor; and iv) upstream intermediates ribose-5-phosphate (measured together with its isomer xyulose-5-phosphate), and 6-phosphogluconate (6PG) [20,25]. We observed an expected reduction in S7P abundance, indicating direct enzymatic loss-of-function in both Shpk^−/− strains (Fig. 2B). We also observed reduction of all other analytes compared to WT and Ctns^−/− (Fig. 2B). Furthermore, as altered Shpk expression corresponded with imbalances in cellular redox states [20,25], we also examined in Shpk^−/− liver the status of energy carriers (Fig. S2A), hexose sugars (Fig. S2B) or acetyl coA (Fig. S2C). With the exception of a decrease of hexose sugars (glucose and fructose) and AMP in Shpk^−/− samples, no significant changes were detected between the different groups. These findings indicate that in vivo elimination of SHPK yields a phenotype of PPP dysregulation in liver without marked changes in overall cellular energy balance.

Fig. 2.

Elimination of Shpk leads to metabolic perturbations in the pentose phosphate pathway. (A) Schematic of the pentose phosphate pathway (PPP) illustrating reactions, intermediates and SHPK metabolism of implicated metabolites. (B) Mass spectrometry quantitation of PPP metabolite relative abundance in WT, Ctns^−/−, Shpk^ΔATG, and Shpk^ΔE2 livers (n = 4–6). (C) Mass spectrometry quantitation of urinary PPP metabolic content (n = 4–6). Statistical significance was determined using one-way ANOVA followed by Tukey’s multiple comparison test. Data are mean ± sem * P < 0.05, ** P < 0.01, *** P < 0.001.

Clinical analysis of human cystinosis patients homozygous for the 57-kb deletion revealed urinary secretion of sedoheptulose and erythritol [19]. Thus, we examined PPP associated metabolites (sedoheptulose, erythronic acid, and erythritol) in urine from Shpk-deficient mice. We detected significant increases in urinary sedoheptulose and erythritol in Shpk^−/− mice as compared to either WT and Ctns^−/− mice (Fig. 2C). These results demonstrate the first in vivo models of PPP dysregulation. The new mouse models appear to successfully phenocopy understudied aspects of metabolic dysregulation in cystinosis, and further highlights on as-yet-unknown aspects of more global perturbation of the PPP in tissues deficient for SHPK.

3.3. Shpk^−/− HSPC transplantation reduced cystine level and restored Ctns expression across multiple tissues of recipient cystinotic mice

Following generation and characterization of Shpk^−/− mice, we evaluated the therapeutic efficacy of Shpk^−/−-HSPC transplantation for cystinosis disease prevention to predict if cystinosis patient homozygous for the 57-kb deletion would benefit from the stem cell gene therapy. Donor Sca1⁺ HSPCs were immunomagnetically harvested from the bone marrow of 6–8-week-old Shpk^ΔATG2 and Shpk^ΔE2 mice and transplanted into 2-month-old lethally irradiated Ctns^−/− mice. As positive, we also transplanted WT-HSPCs in Ctns^−/− mice [9,12,13]. As negative controls, we used Ctns^−/− mice untreated or transplanted with Ctns^−/−-HSPCs; no significant difference existing between these two groups for cystine and Ctns expression analysis (data not shown). As both Shpk^−/− mouse models resulted from a complete knockout of SHPK expression and in similar metabolic phenotype (Figs. 1C, 2 and S2) and because there was no significant difference between these two groups for cystine and Ctns expression analysis (data not shown), recipients of Shpk^ΔATG2 and Shpk^ΔE2 HSPCs were pooled for downstream analysis.

Proper bone marrow reconstitution and hematopoiesis was obtained in the Ctns^−/− mice transplanted by Ctns^−/−, WT and Shpk^−/−-HSPCs as evidenced by complete blood counts (Table S2). Effectiveness of HSPC transplantation for cystinosis can best be assessed by quantifying cystine content across various body tissues. In liver, spleen, muscle, brain, and eye, cystine level was significantly reduced in both WT- and Shpk^−/−-HSPC-transplanted Ctns^−/− mice compared to Ctns^−/− controls (Fig. 3A). As sex-specific differences in cystine accumulation have been observed specifically in the kidney [9], kidney cystine content data were analyzed in males and females separately, cystine was reduced in both male and female Shpk^−/−-HSPC-transplanted Ctns^−/− kidneys compared to controls but significant only in females (Fig. 3A). An additional indication of transplantation efficacy is Ctns mRNA expression, which is normally absent from Ctns^−/− animals. Ctns mRNA expression was measured in the liver, spleen, brain, and kidney of Ctns^−/− mice in the different groups. As compared to Ctns^−/− controls, mice injected with Shpk^−/−-HSPCs had significantly higher Ctns expression across all tissues at a level comparable to WT-HSPC recipients (Fig. 3B). Altogether, these data demonstrate that Shpk^−/−-HSPCs are equally efficacious as WT-HSPCs at exogenous restoration of Ctns expression and subsequent reduction of cystine accumulation.

Fig. 3.

Transplantation of Shpk-deficient HSPCs into cystinosis mice effectively reduced cystine accumulation and restored Ctns expression. (A) Tissue cystine content in Ctns^−/− mice transplanted with WT- and Shpk^−/−-HSPCs compared to Ctns^−/− controls. (B) Ctns mRNA expression across several tissues in Ctns^−/− mice transplanted with WT- and Shpk^−/−-HSPCs compared to Ctns^−/− controls. Data are mean ± sem. * P < 0.05, ** P < 0.01, *** P < 0.001 relative to Ctns^−/− controls.

3.4. Shpk^−/− HSPC transplantation improves kidney function and morphology in Ctns^−/− mice

It was previously shown that WT-HSPCs were able to preserve kidney function and structure [13]. In the past, C57BL/6 Ctns^−/− robustly demonstrated Fanconi syndrome markers such as glucosuria, phosphaturia, proteinuria and polyuria, and developed chronic kidney disease [26]. However, a steady decrease of symptomatic severity of kidney defects has been observed by multiple independent investigators using the same mouse model, potentially due to genetic drift. Thus, as expected, no major difference was observed between WT and Ctns^−/− mice for kidney function (Table 1). Only polyuria was significantly increased in Ctns^−/− mice compared to WT mice, and was significantly reduced in Ctns^−/−-, WT- and Shpk^−/−-HSPCs transplanted Ctns^−/− mice (Table 1).

Renal architecture was evaluated by histological analysis scoring in blind with scores ranging from 1 to 6 based on the percent extent of cortical damage as previously described [13]. Irradiation causes tissue anomalies so untreated and Ctns^−/−-HSPC Ctns^−/− recipients were evaluated separately. The severity of kidney damage was significantly lower in WT-HSPC and Shpk^−/−-HSPC Ctns^−/− recipients as compared to Ctns^−/−-HSPCs Ctns^−/− mice (Fig. 4). The main anomalies consisted in glomerulosclerosis, dilated or atrophic convoluted proximal tubules, thickening of the basement membranes, protein casts, and mononuclear infiltrates, which were seen predominantly in the Ctns^−/−-HSPC Ctns^−/− mice (Fig. 4). Therefore, both WT-HSPCs and Shpk^−/−-HSPCs conferred a protective effect on kidney morphological structure in Ctns^−/− mice.

Fig. 4.

Transplantation of Shpk-deficient HSPCs into Ctns^−/− mice preserves kidney morphology. Representative kidney sections stained with hematoxylin & eosin of 8-month-old Ctns^−/− mice that received either WT-HSPCs, Shpk^−/−-HSPCs or Ctns^−/−-HSPCs. Samples were graded in blind based on cortical damage – see Methods for details. Glomerulosclerosis (arrow), protein casts (arrowhead) and mononuclear infiltrates (^) could be observed especially in the Ctns^−/−-HSPC recipients. Statistical significance was determined using one-way ANOVA followed by Tukey’s multiple comparison test. Data are mean ± sem * P < 0.05.

3.5. Transplanted Shpk^−/−-HSPC progeny are capable of tissue integration and differentiation into macrophages

Delivery of functional cystinosin to diseased cells relies on tissue integration of HSPC-derived macrophages. To characterize HSPC-derived cells in vivo, we generated DsRed⁺Shpk^−/− mice and isolated HSPCs for visualization and characterization of their progeny within tissues after transplantation in Ctns^−/− mice. As control, DsRed⁺WT⁻-HSPCs were transplanted in Ctns^−/− mice. In liver and kidney, DsRed-positive cells co-colocalized with the macrophage markers F4/80 and MHCII in Shpk^−/−-HSPC recipient mice, identifying these cells as macrophages (Fig. 5) as previously shown for WT-HSPCs [9,11–13,15]. MHCII being a marker of M1-like macrophages, theses data suggest that Shpk^−/−-HPSCs differentiate into a typical pro-inflammatory macrophage phenotype as we previously described for cystinosis [15]. These data indicate that Shpk-deficiency does not impact HSPC-derived macrophage differentiation following transplantation in Ctns^−/− mice.

Fig. 5.

Shpk-deficient HSPCs retain the ability to differentiate into macrophages after transplantation in liver and kidney. (A and B) Representative confocal images showing co-localization between transplanted DsRed⁺Shpk ^−/−-HSPCs (red) and the macrophage markers MHCII (A) or F4/80 (B) in recipient Ctns^−/− livers (A) and kidney (B). F-actin visualized via phalloidin staining is shown in purple, and the nuclei stained with dapi is shown in blue. Scale bars, 100 μm.

4. Discussion

The present study characterized the first murine models of Shpk knockout, and the potential impact of the absence of SHPK on the HSPC transplantation therapy for cystinosis. This question has direct relevance for the ongoing clinical trial for cystinosis as a large fraction of human cystinosis patients are homozygous for the 57-kb deletion leading to the lack of SHPK in addition to CTNS gene. Indeed, HSPC therapy is mediated by differentiation of HSPCs into macrophages [14], and SHPK has been shown to be involved in the regulation of macrophage polarization [20].

To investigate if Shpk-deficient HSPCs were still able to rescue cystinosis, we first generated two new mouse models knockout for Shpk with deletions at two different loci of the murine Shpk; absence of SHPK protein was confirmed in both models. The mice did not present any obvious phenotype, breeding with mendelian ratio and exhibiting normal growth and behavior. SHPK/CARKL transcripts were previously detected mainly in liver, kidney, and pancreas, with weaker expression in heart, placenta, brain, and lung [16]. While Shpk mRNA expression was observed in kidney, liver, muscle, heart, spleen and brain in our WT and Ctns^−/− mice, we only detected SHPK protein expression in the liver and kidney. No histological anomalies were found in kidney and liver of the two Shpk^−/− mouse models. In contrast, perturbation of the Pentose Phosphate Pathway (PPP) was demonstrated in both mouse models. SHPK is part of the carbohydrate kinase family that phosphorylate sugars as they enter the cells to prevent their efflux [16]. SHPK enzyme is involved in the phosphorylation of sedoheptulose into sedoheptulose-7-phosphate (S7P) [19], which is part of the PPP, and, as expected, the Shpk^−/− mice exhibited reduction of S7P.

It was unclear if absence of SHPK in cystinosis patients led to any pathologic phenotype. A small study reported no obvious clinical difference between patients carrying mutations causing infantile cystinosis and the ones carrying the 57-kb deletion [27]. However, more recently, two unrelated patients homozygous for nonsense mutation in SHPK were reported with strong but non-overlapping clinical phenotype so the attribution of SHPK was not conclusive [28]. Because no obvious phenotype or histological defects were found in the Shpk^−/− mouse models, our results suggest that no major consequences should result from its deletion in patients carrying the 57-kb deletion at the homozygote state. One particularly interesting finding is that our mouse models effectively recapitulate urinary accumulation of the PPP metabolites erythritol and sedoheptulose seen in the patients affected with cystinosis [19]. While sedoheptulose accumulation made sense to the authors, they questioned the relevance of elevated erythritol with SHPK deficiency [19]. Our findings demonstrated that only Shpk, but not Ctns, knockout mice demonstrated elevated urinary sedoheptulose and erythritol providing further independent evidence that the loss of SHPK itself contributes to this phenotype in cystinosis [28]. However, the mechanism leading to increase erythritol secretion was unclear. A potential explanation was reported as sedoheptulose accumulation results in its conversion to S1P by fructokinase that would be converted to erythrose and then reduced to erythritol [23]. In addition, we showed that PPP metabolite perturbation was also present in the liver of Shpk^−/− mice, suggesting that this pathway is also dysregulated in tissues in cystinosis patients in addition to urine and bloodspots.

The plasticity of macrophage polarization allows them to fulfill numerous biological functions ranging from pro-inflammatory roles classified as either pro-inflammatory mediator of a type I immune response (M1 macrophages), to tissue repair, homeostasis and resolution of inflammation (M2 macrophages) [29,30]. In the context of our stem cell trial for cystinosis, an important function of SHPK was reported by Haschemi and colleagues on its involvement in macrophage polarization [20]. Following lipopolysaccharide (LPS)-induced pro-inflammatory stimulation, SHPK-overexpressing in murine cultured macrophages showed decreased bio-abundance of PPP intermediates G3P, and R5P/X5P while unstimulated SHPK-knockdown macrophages showed conversely increased levels, as well as reduced S7P levels [20]. SHPK-knockdown promoted differentiation of the macrophages towards an M1-like phenotype. We showed previously that in vitro M1-like macrophages have decreased ability to generate tunneling nanotubes involved in lysosomal cross-correction in the hematopoietic stem cell-mediated therapy in cystinosis as opposed to M2-like macrophages [15]. However, in vivo we observed that transplantation of HSPCs isolated from Rac2^−/− mice in which M2-like differentiation is disrupted still resulted in therapeutic efficacy in Ctns^−/− mice, suggesting complex influence of the environmental milieu in macrophage differentiation. In the present study, in contrast to the results obtained in cultured macrophages [20], our Shpk^−/− mouse models yielded a general reduction in bio-abundance of R5P as well as both upstream and downstream PPP intermediates 6PG, E4P, and S7P in the liver without external LPS stimulation. 6PG and E4P were not measured in the previous study [20], so it is not clear to what extent the PPP is different between these two studies, but as Shpk has been implicated with regulating glucose metabolism and regulating metabolic state [20, 28, 31], it is not surprising that Shpk^−/− may result in a more global perturbation of the PPP pathway beyond affecting S7P, though the mechanism is still undefined.

In the limit of the current Ctns^−/− mouse model that presents a milder phenotype than human, this study allowed us through a variety of metrics including HSPC tissue integration and differentiation, kidney histology, tissue Ctns expression, and tissue cystine levels, to confirm that Ctns^−/− recipients of Shpk^−/−-HSPCs showed significant improvements of diseased conditions. Altogether, these data strongly suggest that macrophage differentiation towards a pro- or anti-inflammatory response is dictated by the environmental milieu that may overcome genetic perturbation of their polarization. Given the well-known difficulty of macrophage polarization in culture to translate to the more complex in vivo environment, our data highlights the importance of in vivo exploration of in vitro findings to better represent physiological complexity.

In conclusion, this study reports the first characterization of the Shpk^−/− mice recapitulating the urinary excretion of sedoheptulose and erythritol found in cystinosis patients homozygous for the 57-kb deletion, and suggesting that the PPP metabolic perturbation is also present in their tissues but with minimal consequences in tissue function and architecture. Our results also highlight the importance of in vivo models in the study of macrophage polarization that may respond to complex cell milieu that cannot be reproduced in vitro. As such, we showed that Shpk-deficient HSPCs will likely offer therapeutic efficacy in cystinosis patients carrying the 57-kb deletion at the homozygote state so these patients can now be enrolled in the ongoing stem cell gene therapy clinical trial for cystinosis, while careful comparison of the outcomes of this patient population will be performed with the rest of the enrolled subjects.