Chitotriosidase as a Novel Biomarker for Therapeutic Monitoring of Nephropathic Cystinosis

Significance Statement

The rare, inheritable, lysosomal storage disorder nephropathic cystinosis is caused by mutations in the gene encoding cystinosin, a lysosomal cystine/proton cotransporter. Cystinosis is characterized by accumulation of cystine in all tissues and the development of CKD and multiple extrarenal complications, and is treated with cysteamine, a cystine-depleting agent. Treatment monitoring involves measuring white blood cell cystine levels, but this assay has important limitations and is not available in many countries. In a prospective study of 57 patients with nephropathic cystinosis, the authors demonstrated that chitotriosidase enzyme activity, a marker of macrophage activation, is a significant predictor for adherence to cysteamine therapy and for the presence of extrarenal complications. Their findings suggest that chitotriosidase holds promise as a novel biomarker for monitoring cysteamine treatment and highlight inflammation’s role in cystinosis pathophysiology.

Abstract

Background

Nephropathic cystinosis, a hereditary lysosomal storage disorder caused by dysfunction of the lysosomal cotransporter cystinosin, leads to cystine accumulation and cellular damage in various organs, particularly in the kidney. Close therapeutic monitoring of cysteamine, the only available disease-modifying treatment, is recommended. White blood cell cystine concentration is the current gold standard for therapeutic monitoring, but the assay is technically demanding and is available only on a limited basis. Because macrophage-mediated inflammation plays an important role in the pathogenesis of cystinosis, biomarkers of macrophage activation could have potential for the therapeutic monitoring of cystinosis.

Methods

We conducted a 2-year prospective, longitudinal study in which 61 patients with cystinosis who were receiving cysteamine therapy were recruited from three European reference centers. Each regular care visit included measuring four biomarkers of macrophage activation: IL-1β, IL-6, IL-18, and chitotriosidase enzyme activity.

Results

A multivariate linear regression analysis of the longitudinal data for 57 analyzable patients found chitotriosidase enzyme activity and IL-6 to be significant independent predictors for white blood cell cystine levels in patients of all ages with cystinosis; a receiver operating characteristic analysis ranked chitotriosidase as superior to IL-6 in distinguishing good from poor therapeutic control (on the basis of white blood cell cystine levels of <2 nmol 1/2 cystine/mg protein or ≥2 nmol 1/2 cystine/mg protein, respectively). Moreover, in patients with at least one extrarenal complication, chitotriosidase significantly correlated with the number of extrarenal complications and was superior to white blood cell cystine levels in predicting the presence of multiple extrarenal complications.

Conclusions

Chitotriosidase enzyme activity holds promise as a biomarker for use in therapeutic monitoring of nephropathic cystinosis.

Nephropathic cystinosis (Mendelian Inheritence in Men 219800) is an autosomal recessive lysosomal storage disorder caused by biallelic mutations in the CTNS gene, which codes for the lysosomal cystine transporter cystinosin.¹ Given CTNS is ubiquitously expressed, the absence or malfunction of cystinosin leads to the accumulation and subsequent crystallization of cystine in the lysosomes in all tissues throughout the body.^1–3 The earliest and most severely affected organ is the kidney, which manifests with a severe proximal tubular dysfunction (renal Fanconi syndrome) from the first year of life. Later on, glomerular function is affected and ESKD develops during late childhood or early adolescence.⁴ Also, extrarenal complications (ERCs) develop, mainly affecting the eyes (corneal cystine crystals, retinopathy), endocrine system (primary hypothyroidism, insulin-dependent diabetes mellitus), and neuromuscular system (peripheral myopathy, swallowing dysfunction).¹ Cysteamine, the only available disease-modifying agent, is the cornerstone of treatment of cystinosis.^5,6 A strict therapeutic-monitoring regimen is needed to ensure compliance and efficiency, given that the onset and severity of cystinosis complications are greatly dependent on adherence to optimal dosage.^7,8 Currently, cystine-depleting therapy is monitored via the white blood cell (WBC) cystine level assay. However, the common methodologies for cystine measurement (HPLC or liquid chromatography–tandem mass spectrometry) are technically demanding and expensive, such that the assay is not available in most developing countries,^9,10 and it suffers from inherent impracticalities and difficulties in sampling and storage.^9–11 In addition, because neutrophils, the main cystine-accumulating cells in blood, have a very short lifespan (<24 hours),¹² cystine measurement in WBC represents only a very short period of compliance or therapeutic efficiency.¹¹ WBC cystine levels may not be representative of the total body tissue cystine load, and in case of future treatment modalities, such as hematopoietic stem cell–based gene therapy, WBC cystine levels will be obsolete. In addition, the cut-off value of WBC cystine levels reflecting the adequate therapeutic control remains controversial.

Hence, the search for alternative biomarkers for therapeutic monitoring that are more practical in use, representative of the long-term therapeutic control, and future-treatment proof, is important.¹

Among cells affected by tissue cystine accumulation, macrophages are particularly amenable. Because of their phagocytic nature, they engulf the debris of dead cells in different tissues including the cystine crystals. In addition, lacking the cystine transporter themselves, they are unable to get rid of the accumulated cystine.¹¹ Moreover, macrophages are long-living cells and known to produce considerable amounts of various inflammatory biomarkers in the circulation when activated. Recently, Prencipe et al.¹³ demonstrated inflammasome activation by cystine crystals, and, accordingly, showed significantly elevated levels of inflammasome-dependent cytokines in the plasma of patients with cystinosis. Elmonem et al.¹¹ further identified the significant elevation of plasma chitotriosidase activity, an enzyme produced by activated macrophages, in a cross-sectional cohort of patients with cystinosis compared with both healthy and pathologic renal control groups. The reference values that were determined on the healthy control population in this study, demonstrate age-related differences between children and adults, which are relevant for the interpretation and assessment of chitotriosidase as a diagnostic test.¹¹ In addition, CKD is widely recognized as a condition of chronic inflammation. Importantly, the role of immunosuppressive (IS) treatment in patients receiving kidney transplant (KTx) should be taken into account because IS agents, including corticosteroids, antimetabolites, and calcineurin inhibitors, have shown to affect the activation and function of macrophages.^14–18

Here, we report the results of the first longitudinal, prospective study designed to investigate the potential of four biomarkers of macrophage activation (IL-1β, IL-6, IL-18, and plasma chitotriosidase enzyme activity) as therapeutic monitors for nephropathic cystinosis.

Methods

Patients

A prospective, international, multicenter study was set up in which 61 nephropathic patients with cystinosis on cysteamine therapy were sequentially recruited during the period of the study (October 2015 to January 2018) from three European cystinosis reference centers (University Hospitals Leuven, Belgium; Radboud University Medical Center Nijmegen, The Netherlands; SüdOstBayern Kliniken Traunstein, Germany). Only patients aged 6 months or older who were receiving cysteamine treatment were eligible for recruitment. At the baseline visit, a detailed medical history was obtained and a multisystem clinical examination was performed. On average, patients were followed up every 3–6 months. At each visit, history and clinical examination were recorded and blood and urine samples were obtained for routine laboratory parameters, including complete blood count, C-reactive protein, serum creatinine, and WBC cystine levels. Elevated WBC count (>15,000/µl below age 3 years, and >10,000/µl above age 3 years) and C-reactive protein (>5 mg/L) were used to exclude patients with ongoing infection. eGFR was calculated on the basis of the Schwartz equation for children and the CKD Epidemiology Collaboration (CKD-EPI) equation for adult patients with cystinosis. Data on the ERCs were obtained from all patients and the status of ERC was expressed as a five-topic score, on the basis of the five most important and widely recognized complications observed in nephropathic cystinosis: (1) primary hypothyroidism, (2) retinopathy, (3) insulin-dependent diabetes mellitus, (4) peripheral myopathy, and (5) swallowing dysfunction (Supplemental Table 1).^19–27 Distal myopathy was defined on the basis of the presence of thenar- and/or intraosseous and/or hypothenar atrophy and clinically reduced grip strength. Swallowing dysfunction was clinically defined on the basis of reported symptoms of difficulties on swallowing liquids and/or solid food (Supplemental Table 1).

Patients carrying a homozygous 24-bp duplication in exon 10 of the CHIT1 gene (three patients), or whose WBC cystine values could not be obtained for the corresponding plasma chitotriosidase activity levels (four patients), were excluded from further statistical analyses.

WBC Cystine Assay

Liquid chromatography–tandem mass spectrometry was used for WBC cystine evaluation in all centers as previously described.¹⁰ The concordance of results between the three measuring laboratories was confirmed by the participation in the European Research Network for Evaluation and Improvement of Screening, Diagnosis and Treatment of Inherited Disorders of Metabolism external quality assurance program for WBC cystine and the successful quantification of eight blinded samples of WBC cystine annually.

Chitotriosidase Enzyme Activity Assay

Plasma chitotriosidase activity was assayed on the basis of the fluorometric method developed by Hollak et al.,²⁸ with minor modifications as previously described.²⁹

Genetic Analyses

The detection of the 24-bp duplication mutation in exon 10 of the CHIT1 gene was performed as described previously.³⁰ CTNS gene analysis was obtained in 45 of the 61 patients.

Cytokines Assays

The following cytokines were assayed in plasma of each patient for each visit during the study: IL-1β, IL-6 (Quantikine; R&D Systems, Minneapolis, MN), and IL-18 (MBL Corporation, Nagoya, Japan). Samples were assayed by a sandwich ELISA technique according to manufacturers’ protocols.

Ethical Approval

This study was approved by the corresponding ethical committees at University Hospitals Leuven, Belgium [reference no. s55514 (ML9454)]; Radboud University Medical Center, Nijmegen, The Netherlands (reference no. 2015–2017); and Kliniken Südostbayern, Traunstein, Germany (reference no. 2016–013). Informed consent was translated to match the native language of each patient and was signed by all recruited patients or their legal guardians in case of minors. The study was conducted according to the Declaration of Helsinki, latest version, concerning studies involving human subjects, and to Good Clinical Practice guidelines.

Statistical Analyses

Statistical analyses were performed using GraphPad Prism [version 8.1.0 (221) for Macintosh] and SAS software (version 9.4 of the SAS System for Windows). Parametric or nonparametric tests were applied depending on the distribution of the data. In the analysis of the longitudinal data for determining predictors of WBC cystine, linear mixed models were applied to account for the repeated measurements per patient. In the correlation analysis between the proposed biomarkers and the number of ERCs in the subgroup of patients harboring at least one ERC, a Spearman correlation coefficient (rho) was calculated for each variable assessed (Supplemental Table 2). Receiver operating characteristic (ROC) curves were created and evaluated using easyROC statistical software.³¹ Data were represented as ratios and percentages for categorical variables and as means and SD or medians and interquartile ranges for numerical variables, unless otherwise stated. Statistical significance was on the basis of 95% confidence intervals (95% CIs) or a P value <0.05 All reported P values are two-sided, unless otherwise stated.

Results

Baseline Demographic, Clinical, and Laboratory Characteristics of Patients with Cystinosis

Demographic, clinical, and laboratory data on all recruited patients with cystinosis are summarized in Table 1. Of the 61 patients with cystinosis, WBC cystine levels were only obtained in 57 patients, because unavailability of access to the WBC cystine assay at the moment of the regular care follow-up visits. Of these 57 patients, 22 were children, five were affected with the juvenile cystinosis phenotype, 26 patients harbored at least one ERC and 30 patients had undergone kidney transplantation before the start of the study (Table 1). One patient was receiving peritoneal dialysis during the full duration of the study.

Table 1.

Demographic, clinical, and laboratory parameters of the cystinosis patient cohort and subgroups stratified according to WBC cystine level < or ≥2 nmol 1/2 cystine/mg protein

Parameter	Unit/Subgroup	All Patients with Cystinosis (n=57)	2-yr Average WBC Cystine Level					Difference (95% CI of the Difference)	P Value
			<2 nmol 1/2 Cystine/mg Protein (n=42)			≥2 nmol 1/2 Cystine/mg Protein (n=15)
Age at recruitment	Yr (mean±SD)	22±11	21±12			24±10		3.3±3.4 (−3.4 to 10)	0.32
Age category at recruitment	Children (<18)/adults (≥18) (% children)	39%	18/24		43%	4/11	27%	16% (−8% to 48%)	0.36
	<12 (%)	19%	10/42		24%	1/14	7%	17% (−0.05% to 46%)	0.26
	12–18 (%)	19%	8/42		19%	3/12	20%	1% (−20% to 31%)	>0.99
	18–30 (%)	42%	16/42		38%	8/7	53%	15% (−15% to 43%)	0.36
	30–40 (%)	12%	5/42		12%	2/13	13%	1% (−17% to 31%)	>0.99
	>40 (%)	7%	3/42		7%	1/14	7%	0% (−14% to 28%)	>0.99
Sex	Male/female (% male)	51%	22/20		52%	7/8	47%	4% (−20% to 29%)	0.77
Phenotype	Infantile/juvenile (% infantile)	91%	38/4		90%	14/1	93%	3% (−25% to 18%)	>0.99
Genetic background	Hom 57 kb del/other (% hom 57 kb del)	48%	13/19		41%	8/4	67%	38% (3% to 6%)	0.02
Age at diagnosis	Yr (mean±SD)	1.7±1.1	1.8±1.2			1.5±0.7		0.3±0.38 (−1.1 to 0.5)	0.44
Kidney transplantation	Yes/no (% yes)	53%	19/23		45%	11/4	73%	28% (−5% to 51%)	0.08
eGFR	nKTx (ml/min per 1.73 m2)	65±30	66±32			59±21		−8±17 (−42 to 27)	0.66
	KTx (ml/min per 1.73 m2)	59±30	70±25			41±29		−29±10 (−50 to −9)	0.007
ERCs	Yes/no (% yes)	46%		19/23	45%	7/8	47%	2% (−28% to 31%)	>0.99
	Primary hypothyroidism	28%		10/32	24%	6/9	40%	16% (−11% to 46%)	0.32
	Retinopathy	4%		1/41	2%	1/14	7%	5% (−9% to 32%)	0.46
	Insulin-dependent Diabetes mellitus	5%		2/40	5%	1/14	7%	2% (−12% to 29%)	>0.99
	Peripheral myopathy	25%		9/33	21%	5/10	33%	12% (−14% to 42%)	0.49
	Swallowing dysfunction	18%		8/34	19%	2/13	13%	6% (−13% to 36%)	>0.99
WBC cystine level	nmol 1/2 cystine/mg protein	1.44 (0.64 to 2.3)		1.12 (0.54 to 1.47)		3.2 (2.56 to 4.85)		2.08 (1.8 to 2.97)	<0.001
	<1 (%)	32%		18/24	43%	N/Aa	N/A
	1–2 (%)	42%		24/18	57%	N/A	N/A
	>2 (%)	26%		N/A	N/A	15/0	100%
Disease severity classification	nKTx, ERC− (% yes)	30%		13/29	31%	4/11	27%	4% (−20% to 36%)	>0.99
	nKTx, ERC+	18%		10/32	24%	0/15	0%	24% (8% to 52%)	0.05
	KTx, ERC−	25%		10/32	24%	4/11	27%	3% (−21% to 33%)	>0.99
	KTx, ERC+	28%		9/33	21%	7/8	47%	25% (−4% to 53%)	0.09
Chitotriosidase	nmol/ml plasma per hour (median [IQR])	77.11 (49.88 to 166.3)		71.33 (43 to 128.5)		160 (69.75 to 445.3)		88.67 (18.83 to 266.3)	0.006
IL-1β	pg/ml (median [IQR])	0.13 (0.13 to 0.19)		0.13 (0.13 to 0.19)		0.13 (0.13 to 0.18)		0 (−0.01 to 0)	0.6
IL-6	pg/ml (median [IQR])	1.68 (1.15 to 2.85)		1.56 (1 to 2.46)		2.08 (1.68 to 4.32)		0.52 (0.09 to 1.72)	0.02
IL-18	pg/ml (median [IQR])	717.0 (499.3 to 939.3)		705.8 (502.9 to 936.5)		747.5 (466.1 to 995.9)		41.69 (−174.4 to 281.8)	0.80

Hom, homozygous; del, deletion; nKTx, non-kidney transplantation; ERC−, no extra-renal complications; ERC+, extra-renal complications; IQR, inter-quartile range.

^aOnly patients of whom WBC cystine levels could be obtained are included in this table. Results for the genetically confirmed enzyme deficient (n=3) were excluded from the analysis for chitotriosidase.

For each patient with cystinosis, two to nine visits were recorded (mean 4.7±2.5), during which samples for macrophage biomarkers were obtained. The average duration of follow-up for all recruited patients was 19.3±6.5 months. None of the patients received kidney graft during the course of the study. Three patients had very low chitotriosidase activity and were confirmed homozygous for the 24-bp duplication of the CHIT1 gene, and hence excluded from further statistical analysis involving chitotriosidase. The reproducibility of chitotriosidase, i.e., the interchangeability of chitotriosidase measurements over time, quantified via determination of the intraclass correlation coefficient, yielded 0.88 (95% CI, 0.82 to 0.92).

Plasma Chitotriosidase Enzyme Activity as a Potential Therapeutic Monitor for Cystine-Depleting Therapy

Forty two patients showed WBC cystine values <2 nmol 1/2 cystine/mg protein on average during the course of the study (73.6%) and were considered as sufficiently controlled, whereas 15 patients showed values >2 nmol 1/2 cystine/mg protein (26.3%), corresponding to poor disease control or insufficient adherence to cystine-depleting therapy (Table 1). Patients with WBC cystine levels < and >2 nmol 1/2 cystine/mg protein did not significantly differ in age, sex, clinical phenotype of cystinosis, or the most important ERCs of cystinosis. In addition, in the subgroup of patients without kidney transplantation, no significant difference in eGFR was observed (Table 1).

Among the four biomarkers of macrophage activation tested, plasma chitotriosidase enzyme activity showed significantly higher levels with increasing WBC cystine levels when considering the average values over the whole 2-year study period (Figure 1A).

Figure 1.

Plasma chitotriosidase enzyme activity levels significantly correlate with the 2-year average WBC cystine levels in the cystinosis patient cohort. (A) Median values for biomarkers of macrophage activation in patients with cystinosis in relation to WBC cystine levels. Biomarker values (IL-1β, IL-6, IL-18, and chitotriosidase) were categorized to WBC cystine values <1, 1–2, 2–3, and ≥3 nmol 1/2 cystine/mg protein. Data are presented as median±IQR. P values were on the basis of Kruskal–Wallis test. (B) Chitotriosidase shows a significant linear correlation with WBC cystine levels in the whole cystinosis patient cohort. The correlation is on the basis of the 2-year average values of chitotriosidase and corresponding WBC cystine levels per patient over the course of the study. The full line represents the estimated linear trend, with the dashed lines representing the 95% CI.

In addition, a significant linear correlation was demonstrated between the 2-year average plasma chitotriosidase activity levels and WBC cystine levels (Figure 1B). Furthermore, to correct for multiple possible confounding covariates, in a mixed-effects multivariate regression analysis in all patients with cystinosis, using linear mixed models to account for the multiple measurements per patient, chitotriosidase activity resulted as a significant predictor for WBC cystine levels, together with age and IL-6 (Table 2).

Table 2.

Mixed-effects multivariate linear regression analysis on the longitudinal data for predictors of WBC cystine level in all patients with cystinosis

Variable	Level	Estimate (95% CI)	P Value
WBC cystine
Age at recruitment		−0.95 (−1.78 to −0.13)	0.02
Sex	Female	−0.48 (−1.04 to 0.07)	0.09
Chitotriosidase		0.005 (0.003 to 0.007)	<0.001
IL-1β		0.22 (−1.14 to 1.59)	0.75
IL-6		0.12 (0.05 to 0.19)	<0.001
IL-18		0.00 (−0.00 to 0.00)	0.56
Genetic background	Hom 57 kb del	0.23 (−0.33 to 0.79)	0.42

Hom, homozygous; del, deletion.

Using ROC analysis, the performance of chitotriosidase and IL-6 was assessed in distinguishing WBC cystine levels < or ≥2 nmol 1/2 cystine/mg protein, corresponding to patients with good versus poorly controlled cystinosis. Herein, chitotriosidase activity was superior to IL-6 (Figure 2).

Figure 2.

Plasma chitotriosidase enzyme activity performs better than other tested biomarkers of macrophage activation, in distinguishing good versus poor therapeutic control (WBC cystine <2 versus ≥2 nmol 1/2 cystine/mg protein). (A–D) Box and whisker plots for IL-1β, IL-6, IL-18, and plasma chitotriosidase enzyme activity levels in all patients with cystinosis, according to the 2-year average WBC cystine level (<2 or ≥2 nmol 1/2 cystine/mg protein). (B) *P=0.0263. (D) ** P=0.006. (E and F) ROC curve analysis of IL-6 and plasma chitotriosidase enzyme activity demonstrating their performance for distinguishing good versus poor therapeutic control. AUC, area under the curve.

WBC Cystine Is a Significant Independent Predictor for Chitotriosidase

Because CKD is a state of general inflammation, and plasma chitotriosidase enzyme activity is influenced by kidney function, we aimed to corroborate our hypothesis that the burden of cystine accumulation, reflected by WBC cystine levels, exert an independent effect on chitotriosidase enzyme activity levels, on top of kidney function (eGFR). In a multivariate regression analysis, we demonstrated that WBC cystine is a significant (P=0.001) independent predictor of chitotriosidase levels in the whole cystinosis patient cohort (Table 3). Hence, chitotriosidase reflects the cystine burden, independent of kidney function.

Table 3.

Mixed-effects univariate (1, 2) and multivariate (3) regression analysis on the longitudinal data for predictors for chitotriosidase in all patients with cystinosis

Variable–effect	Slope (95% CI)a	P Value	AICb
Chitotriosidase
(1) eGFR–linear	−0.98 (−1.71 to −0.25)	0.009	2331.49
(2) WBC cystine–linear	9.38 (4.04 to 14.72)	<0.001	2326.80
(3) Multivariate model: eGFR+WBC cystine			2323.03
eGFR–linear	−0.88 (−1.59 to −0.16)	0.02
WBC cystine–linear	8.85 (3.56 to 14.15)	0.001

^aAkaike Information Criterion: smaller is better.

^bSlope >(<)0: higher (lower) chitotriosidase level for higher leucocyte cystine level.

The Number of IS Agents Used in KTx Patients with Cystinosis Is a Significant Independent Predictor of Chitotriosidase

Furthermore, we aimed to explore the effect of IS treatment in KTx patients with cystinosis.

Overall, KTx patients with cystinosis demonstrated a significantly higher chitotriosidase level compared with non-KTx patients with cystinosis (Table 4). However, KTx patients with cystinosis were significantly older, had higher WBC cystine levels, and tended to have a slightly lower kidney function compared with non-KTx patients (Table 4). Therefore, in a multivariate regression analysis, we aimed to determine whether the number of IS agents used is an independent significant predictor for chitotriosidase, correcting for age, sex, eGFR, and WBC cystine levels (Tables 5 and 6). Indeed, the number of IS agents is a significant independent predictor for chitotriosidase in the KTx population, apart from WBC cystine levels, with the number of immunosuppressants being inversely correlated with chitotriosidase (Table 6). Because of the high number of various IS drugs (n=7) and treatment regimens (n=8) being applied in KTx patients, and the low number of KTx patients per regimen, further elucidation of the relation of certain agents or regimens with chitotriosidase levels was not feasible.

Table 4.

Demographic, clinical, and laboratory parameters of the cystinosis patient cohort and subgroups of KTx versus non-KTx patients

Parameter	Unit/Subgroup	All Patients with Cystinosis (n=57)	Kidney Transplantation Status		Difference (95% CI of the Difference)	P Value
			nKTx	KTx
Age at recruitment	Yr (mean±SD)	22±11	13±7	29±8	15.8±1.9 (12 to 20)	<0.001
Age category at recruitment	Children (<18)/adults (≥18) (% children)	39%	83%	3%	80% (69% to 104%)	<0.001
	<12 (%)	19%	38%	0%	38% (18% to 59%)	<0.001
	12–18 (%)	19%	45%	3%	42% (22% to 65%)	0.002
	18–30 (%)	42%	17%	63%	24% (2% to 41%)	0.03
	30–40 (%)	12%	0%	22%	22% (3% to 40%)	0.02
	>40 (%)	7%	0%	13%	13% (−4% to 30%)	0.12
Sex	Male/female (% male)	51%	45%	50%	5% (−21% to 30%)	0.8
eGFR	ml/min per 1.73 m2	65±30	65±30	59±30	−5.9±7.9 (−21.8 to 10)	0.46
ERCs	Yes/no (% yes)	46%	34%	50%	16% (−11% to 39%)	0.3
	Primary hypothyroidism	28%	14%	38%	24% (−0.9% to 45%)	0.04
	Retinopathy	4%	0%	6%	6% (−9% to 22%)	0.49
	Insulin-dependent diabetes mellitus	5%	0%	10%	9% (−7% to 26%)	0.24
	Peripheral myopathy	25%	14%	31%	17% (−6% to39%)	0.13
	Swallowing dysfunction	18%	7%	25%	18% (−3% to 38%)	0.08
WBC cystine level	nmol 1/2 cystine/mg protein	1.44 (0.64 to 2.3)	1.23 (0.54 to 1.6)	1.54 (0.99 to 3.04)	0.31 (0.04 to 1.18)	0.03
	<1 (%)	32%	41%	23%	17% (−6% to 44%)	0.25
	1–2 (%)	42%	44%	40%	4% (−22% to 31%)	0.79
	>2 (%)	26%	15%	37%	22% (−4% to 44%)	0.08
Chitotriosidase	nmol/ml plasma per hour (median [IQR])	77.11 (49.88 to 166.3)	64.5 (33.8 to 108)	109.8 (70.38 to 317.8)	45.25 (22.25 to 147.1)	<0.001
IL-1β	pg/ml (median [IQR])	0.13 (0.13 to 0.19)	0.13 (0.13 to 0.16)	0.13 (0.13 to 0.19)	0.0042 (0 to 0.02)	0.26
IL-6	pg/ml (median [IQR])	1.73 (1.21 to 2.66)	1.65 (0.98 to 2.16)	1.81 (1.30 to 3.71)	0.16 (−0.13 to 1.09)	0.11
IL-18	pg/ml (median [IQR])	717.0 (499.3 to 939.3)	773.1 (533.9 to 1485)	650.1 (484.8 to 863.3)	−122.9 (−401.4 to 13.89)	0.07

nKTx, non-kidney transplanted; IQR, interquartile range.

Table 5.

Demographic, clinical, and laboratory parameters of the subgroup of KTx patients with cystinosis stratified according to the number of IS agents

Parameter	Unit	KTx Patients (n=26)	No. of IS Agents
			1 (n=2)	2 (n=16)	3 (n=8)	P Value
Age	Yr (mean±SD)	29±8	36±12	29±8	29±6	0.46
Age at KTx	Yr (mean±SD)	14±5	9±1	13±3	16±6	0.11
Time since KTx	Yr (mean±SD)	16±10	27±11	15±10	14±9	0.24
No. of IS agents	Median (IQR)	2 (2–3)	1	2	3
Type of IS
Prednisolone	Y/N (% Y)	23/3 (88%)	0/2 (0%)	15/1 (94%)	8/0 (100%)
CNI	Y/N (% Y)	18/8 (69%)	2/0 (100%)	8/8 (50%)	8/0 (100%)
Cyclosporine	Y/N (% Y)	4/22 (15%)	0/2 (0%)	2/14 (13%)	2/6 (25%)
Tacrolimus	Y/N (% Y)	14/12 (54%)	2/0 (100%)	6/10 (38%)	6/2 (75%)
Antimetabolite	Y/N (% Y)	15/11 (58%)	0/2 (0%)	7/9 (44%)	8/0 (100%)
Azathioprine	Y/N (% Y)	5/21 (19%)		2/7 (22%)	3/5 (38%)
MMF	Y/N (% Y)	10/16 (38%)		5/4 (56%)	5/3 (62%)
mTOR inhibitor	Y/N (% Y)	3/23 (12%)	0/2 (0%)	0/16 (0%)	1/7 (13%)
Everolimus	Y/N (% Y)	2/24 (8%)			1/7 (13%)
Sirolimus	Y/N (% Y)	1/25 (4%)			0/8 (0%)
WBC cystine	nmol 1/2 cystine/mg protein (mean±SD)	2.46±2.4	2.8±3.1	2.7±2.8	1.8±1.1	0.7
eGFR	ml/min per 1.73 m2	58±28	40±37	64±32	52±14	0.4
Chitotriosidase	nmol/ml plasma per hour (median [IQR])	197±189	462±526	204±145	117±125	0.06

IQR, interquartile range; Y, yes; N, no; CNI, calcineurin inhibitor; MMF, mycophenolate mofetil; mTOR, mammalian target of rapamycin.

Table 6.

Multivariate regression analysis for predictors of chitotriosidase, including the number of IS agents, in the subgroup of KTx patients with cystinosis

Variable	Estimate (95% CI)	P Value
Chitotriosidase
No. of IS compounds	−126.0 (−225.6 to −26.41)	0.01
Age	0.016 (−7.653 to 7.69)	0.1
Sex, female	−25.69 (−142.2 to 90.79)	0.66
eGFR	−0.91 (−2.46 to 0.65)	0.25
WBC cystine	10.36 (0.72 to 20.004)	0.04

Chitotriosidase Enzyme Activity as a Surrogate Biomarker for Disease Severity

Twenty six of the 61 patients showed at least one ERC. Relevant demographic and clinical data are depicted in Table 7.

Table 7.

Demographic, clinical, and laboratory parameters of the subgroup of patients with cystinosis with ERCs, according to the presence of one or multiple ERCs

Parameter	Unit/Subgroup	All Patients with Cystinosis (n=57)	ERC Status				Difference (95% CI of the Difference)	P Value
			1 ERC (n=16)		>1 ERC (n=10)
Age at recruitment	Yr (mean±SD)	22±11	21±9		33±9		12±4 (4 to 19)	0.003
Age category at recruitment	Children (<18)/adults (≥18) (% children)	39%	44%		0%		44% (18% to 85%)	0.02
	<12 (% yes)	19%	19%		0%		19% (−9% to 56%)	0.26
	12–18	19%	25%		0%		25% (−3% to 63%)	0.14
	18–30	42%	38%		50%		13% (−28% to 49%)	0.69
	30–40	12%	19%		30%		11% (−24% to 49%)	0.64
	>40	7%	0%		20%		20% (−9% to 56%)	0.14
Sex	Male/female (% male)	51%	63%		60%		3% (−35% to 41%)	>0.99
Phenotype	Infantile/juvenile (% infantile)	91%	100%		100%		∞ (−0.34 to 0.24)	>0.99
Genetic background	Hom 57 kb del/other (% hom 57 kb del)	48%	73%		63%		10% (−31% to 53%)	>0.99
Kidney transplantation	(% Yes)	53%	38%		100%		63% (19% to 84%)	0.003
eGFR	All patients		57.7 (36.8 to 82.9)		47.8 (13.7 to 76.8)		−9.9 (−34.9 to 22.2)	0.44
	nKTx	66 (40 to 86)	79 (37 to 84)		N/A			N/A
	KTx	59 (39 to 82)	38.8 (30 to 81)		48 (14 to 77)		9 (−35 to 34)	0.87
ERCs	Yes/no (% yes)	46%	16/0	100%	10/0	100%
	Primary hypothyroidism	28%	6/10	38%	10/0	100%	63% (19% to 84%)	0.003
	Retinopathy	4%	0/16	0%	2/8	20%	20% (−9% to 56%)	0.14
	Insulin-dependent diabetes mellitus	5%	0/16	0%	3/7	30%	30% (−3% to 65%)	0.05
	Peripheral myopathy	25%	7/9	44%	7/3	70%	26% (−17% to 58%)	0.25
	Swallowing dysfunction	18%	3/13	19%	73	70%	51% (7% to 77%)	0.02
WBC cystine level	nmol 1/2 cystine/mg protein	1.44 (0.64 to 2.3)	1.12 (0.54 to 1.78)		2.43 (1.38 to 4.88)		1.31 (0.06 to 3.53)	0.03
	<1 (% yes)	32%	7/9	44%	2/8	20%	24% (−7% to 66%)	0.4
	1–2	42%	7/9	44%	3/7	30%	14% (−20% to 55%)	0.68
	>2	26%	2/14	13%	5/5	50%	38% (−3% to 69%)	0.07
Chitotriosidase	nmol/ml plasma per hour (median [IQR])	77.11 (49.88 to 166.3)	54 (32 to 104)		321.8 (78.1 to 495.7)		267.8 (36 to 416.5)	0.005
IL-1β	pg/ml (median [IQR])	0.13 (0.13 to 0.19)	0.13 (0.13 to 0.19)		0.13 (0.13 to 0.19)		0 (−0.03 to 0.03)	0.95
IL-6	pg/ml (median [IQR])	1.68 (1.15 to 2.85)	1.65 (1.14 to 3.16)		2.9 (1.68 to 4.71)		1.2 (−0.2 to 2.5)	0.07
IL-18	pg/ml (median [IQR])	717.0 (499.3 to 939.3)	813.2±485		719±182.9		−94.17±161 (−426.5 to 238.1)	0.56

Hom, homozygous; del, deletion.

Only chitotriosidase activity, apart from WBC cystine levels, yielded a significant correlation with the number of ERCs (Supplemental Table 2, Supplemental Figure 1). Upon grouping patients with one versus multiple ERCs, a significant difference was observed in age at recruitment, the status of kidney transplantation, the presence of specific ERCs including primary hypothyroidism and swallowing dysfunction, the average WBC cystine level, and the average chitotriosidase activity level (Table 7). In an ROC analysis, the performance of plasma chitotriosidase enzyme activity for distinguishing one versus multiple ERCs was superior (area under the curve [AUC] chitotriosidase, 0.83; 95% CI, 0.64 to 1.01) to WBC cystine levels (AUC WBC cystine, 0.75; 95% CI, 0.53 to 0.96) (Figure 3, Table 8).

Figure 3.

Plasma chitotriosidase enzyme activity performs better than WBC cystine levels and other tested biomarkers of macrophage activation, in distinguishing the presence of one versus multiple ERCs in patients harboring at least one ERC. (A and B) Box and whisker plots for the 2-year average WBC cystine level and plasma chitotriosidase enzyme activity levels in patients harboring at least one ERC, stratified according to the presence of one versus multiple ERCs. (A) *P=0.04. (B) **P=0.006. (C and D) ROC curve analysis of 2-year average WBC cystine and plasma chitotriosidase enzyme activity levels demonstrating the superior performance of chitotriosidase for distinguishing the presence of multiple versus a single ERC. AUC, area under the curve.

Table 8.

Cut-off levels of chitotriosidase for predicting good versus poor therapeutic control, and the presence of one versus multiple ERCs and its corresponding performance measures determined by ROC analysis

Biomarker	Population	Cut-Off	Sensitivity	Specificity	PPV	NPV	+LR	−LR
Good versus poor therapeutic control
Chitotriosidase (nmol/ml plasma per hour)	All patients	150	53% (30% to 75%)	85% (70% to 93%)	57% (33% to 79%)	83% (68% to 91%)	1.24	0.55
Presence of one versus multiple cystinosis ERCs
Chitotriosidase (nmol/ml plasma per hour)	ERC+ patients	250	60% (31% to 83%)	93% (70% to 100%)	86% (49% to 99%)	78% (55% to 91%)	4.2	0.42

PPV, positive predictive value; NPV, negative predictive value; +LR, positive likelihood ratio; −LR, negative likelihood ratio.

Taken together, although a cut-off value for chitotriosidase enzyme activity of 150 nmol/ml plasma per hour resulted in a specificity of 85% and a negative predictive value of 83% for distinguishing good versus poor therapeutic control (Figure 4, Table 8)m a cut-off value for chitotriosidase enzyme activity of 250 nmol/ml plasma per hour resulted in a specificity of 93% and a positive predictive value of 86% in identifying patients harboring multiple ERCs (Figure 4, Table 8).

Figure 4.

Plasma chitotriosidase enzyme activity harbors a high negative predictive value for ruling out insufficient cysteamine treatment and a high positive predictive value for identifying patients harboring multiple ERCs. (A) In the whole cystinosis patient cohort, a plasma chitotriosidase enzyme activity level <150 nmol/ml plasma per hour has a high negative predictive value for ruling out poor compliance to cysteamine treatment, on the basis of a high (2-year) average WBC cystine level ≥2 nmol 1/2 cystine/mg protein. (B) In patients with cystinosis who are harboring at least one ERC, a plasma chitotriosidase enzyme activity level ≥250 nmol/ml plasma per hour has a high positive predictive value for the presence of multiple ERCs.

Discussion

We conducted a 2-year longitudinal study to explore the clinical value of biomarkers of macrophage activation as potential additional monitors for cystine-depleting therapy in nephropathic cystinosis.

The rationale for this study was that WBC cystine measurements are not available in the majority of the countries and are subject of large variability even in the reference laboratories. On the other hand, WBC cystine levels may not adequately reflect the whole-body cystine burden and long-term compliance. Alternatively, interstitial macrophages containing cystine crystals have been demonstrated in various organs, including the skin, gastrointestinal mucosa, liver, kidney, and bone marrow,^32–37 and macrophage activation has been established as one of the important pathogenic mechanisms of cystinosis.^11,13,38 In addition, substantial evidence indicates a pivotal role for inflammation mediated by macrophages, comprising mainly the macrophage-secreted cytokines IL-6 and TNF-α, the M1/M2 macrophage balance, and macrophage-derived monocytes in the progression of CKD toward ESKD.^39–46

Therefore, we hypothesized that widespread progressive cystine crystal accumulation in various tissues can be reflected by inflammatory mediators that are released by macrophages upon exposure to cystine crystals. Likewise, the degree of inflammation mediated by macrophages could therefore reflect the long-term adherence to cystine-depleting therapy.^34,35

Among the four biomarkers for macrophage activation investigated, plasma chitotriosidase enzyme activity was the only promising candidate for clinical use as an additional monitor for therapeutic monitoring of cystinosis.

Plasma chitotriosidase enzyme activity correlated significantly with WBC cystine values over the longitudinal course of the study, and resulted as a significant predictor for WBC cystine levels in patients of all ages with cystinosis.

Importantly, we demonstrated that WBC cystine level is an independent significant predictor for chitotriosidase, acting on top of kidney function (eGFR). Hence, chitotriosidase does not merely reflect the inflammation related to CKD in the cystinosis population, which suggests that it can be used to evaluate whole-body cystine burden. Although the optimal WBC cystine level reflecting adequate cysteamine treatment is a matter of controversy, we have defined the cut-off for distinguishing appropriate versus poor disease control at 2 nmol 1/2 cystine/mg protein for various arguments.^47–50

The value of 1 nmol 1/2 cystine/mg protein has mainly been supported by only one retrospective report, in which a better preservation of kidney function was observed in a subgroup of patients with median mixed leukocyte cystine levels <1.0 nmol 1/2 cystine/mg protein.⁴⁸ In contrast, in the study by Markello et al.,⁴⁹ the subgroup of patients with cystinosis that was defined to be adequately treated showed average WBC cystine levels >1 nmol 1/2 cystine/mg protein (1.1±0.7). In addition, in the study by Gahl et al.⁵⁰ adequate cystine depletion was defined as WBC cystine levels <2.5 nmol 1/2 cystine/mg protein. In our clinical experience, WBC cystine levels <1 nmol 1/2 cystine/mg protein are very difficult to achieve, especially in adolescent and adult patients, because of the side effects of cysteamine and the frequent dosing regimen (four times a day) of the most widely available formulation of immediate release cysteamine. In our cohort, only 18 of all patients with cystinosis showed an average 2-year WBC cystine level <1 nmol 1/2 cystine/mg protein. Finally, it is uncertain whether the difference between a WBC cystine <1 and <2 nmol 1/2 cystine/mg protein is clinically relevant. In an ROC curve analysis, the AUC of plasma chitotriosidase enzyme activity levels did not allow to separate patients with a WBC cystine level ≥1 versus <1 nmol 1/2 cystine/mg protein, indicating that these two populations may have a similar cystine burden (data not shown). On the other hand, only six patients showed an average 2-year WBC cystine level ≥3, which is insufficient to yield a properly powered ROC curve analysis for assessing the value of chitotriosidase in distinguishing WBC cystine levels ≥ or <3 nmol 1/2 cystine/mg protein.

Interestingly, in the KTx patients with cystinosis, the number of IS agents used is a significant independent predictor for chitotriosidase, apart from WBC cystine levels, pointing to the fact that chitotriosidase reflects the general state of inflammation. These results are in line with several other reports on the effect of corticosteroids, antimetabolites, and calcineurin inhibitors on macrophage function.^14–18

On the other hand, the low number of patients in the subgroups of number of IS agents hampers firm conclusions on the effect of level of immunosuppression on chitotriosidase enzyme activity in this population, and needs to be further elucidated.

In addition, of all biomarkers for macrophage activation tested, chitotriosidase enzyme activity was the only biomarker that demonstrated a significant correlation with the number of ERCs in the subgroup of patients harboring at least one ERC. Furthermore, in an ROC analysis, chitotriosidase activity yielded a superior performance to WBC cystine levels for identifying patients harboring multiple ERCs for which a cut-off value could be established yielding a high positive predictive value (86%). Hence, our data indicates that plasma chitotriosidase enzyme activity can reflect the severity and extent of disease, as far as the clinical renal and extrarenal manifestations are mediated by cystine crystal accumulation and its induced inflammation.

Conversely, the highest chitotriosidase levels are found in the older, adult KTx patients with high WBC cystine levels and multiple ERCs, and thus reflect more seriously affected patients. Hence, it can be speculated that these patients have undergone a period of poor adherence to cysteamine treatment, and thus chitotriosidase could hold the potential for a biomarker reflecting long-term adherence to cystine-depleting therapy.

Initially, chitotriosidase enzyme was suggested as an important factor in the innate immunity against chitin-coated pathogens. However, its immunomodulatory effects extend far beyond innate immunity.⁵¹ Chitotriosidase was first detected as markedly elevated in the plasma of patients with Gaucher disease,²⁸ and later it was found to be elevated in patients with other lysosomal storage disorders such as Niemann-Pick A/B and C, Gangliosidosis M1, and Krabbe,^52,53 as well as inflammatory disorders in which macrophages play an important role, such as β-thalassemia, sarcoidosis, and multiple sclerosis.^54–56 The enzyme is currently established as a therapeutic monitor for Gaucher disease.⁵⁷ In cystinosis, we detected the elevation of the enzyme activity in the plasma of cystinosis knocked-out mice,^11,58 and in the homogenates of the cystinosis mutant zebrafish larvae compared with the wild type in each animal model.⁵⁹

Chitotriosidase activity has many advantages as a biomarker. The enzyme is stable in plasma for a very long time (1 month at room temperature, >4 months at 4°C, and several years at –80°C) and retains its stability upon repeated freezing and thawing for up to ten cycles.^29,60 Although a novel methodology by which granulocytes can be specifically and quickly isolated using immunomagnetic purification, allowing for a 30-hour time window for shipping of samples to specialized laboratories, has partly resolved the issue of quick processing, the stability of chitotriosidase allows substantially more versatility.⁶¹

Analysis of chitotriosidase enzyme activity is performed through a simple fluorometric technique at a fraction of the time and cost of WBC cystine assay, and could be easily available to much less equipped laboratories in developing countries. Furthermore, unlike cystine the enzyme can be assayed in dried blood spots, providing another sample type that has many advantages concerning sample storage and transport.²⁹

Importantly, our previous cross-sectional study showed a correlation between chitotriosidase enzyme activity and kidney function, with highest levels observed in patients with poorly controlled cystinosis in advanced stages of CKD.¹¹ Higher chitotriosidase activity levels were also observed in other patients with CKD, reflecting a general state of inflammation.¹¹ The levels in cystinosis were, however, significantly higher compared with noncystinotic patients with a similar CKD stage.¹¹

Although promising, our study has some limitations. The 24-bp duplication in exon 10 of the chitotriosidase gene (CHIT1), when homozygously mutated, results in chitotriosidase deficiency. It is relatively common in many ethnic groups, including whites (approximately 5% of the general population).⁶² In our study, we detected very low activities of the enzyme associated with homozygosity for the 24-bp duplication in three of our recruited patients, coinciding with the mutation prevalence. Naturally, the enzyme in those patients cannot be used as a therapeutic monitor for cystinosis therapy in the future. When suspected, this mutation can be detected with a PCR reaction, and was not an obstacle its clinical use as a biomarker in other diseases.⁵²

Furthermore, cystinosis is a very rare disease. Although we were able to follow 61 cystinosis patient over a time period of 2 years, this group was still small, especially given that the number of patients with poor disease control was less than one fourth of all patients. Moreover, patients with advanced stages of CKD, and especially patients on dialysis, were underrepresented in our population. Hence, the usefulness of chitotriosidase in patients with cystinosis who are receiving dialysis needs to be further explored.

Nevertheless, even with these low numbers we were able to demonstrate that chitotriosidase enzyme activity levels <150 nmol/ml per hour have a high predictive value for ruling out poor adherence to cysteamine therapy in all patients with cystinosis, which can have a direct therapeutic implication for adjusting cysteamine therapy.

Obviously, a similar study in a larger patient population, including a significant subgroup of patients on dialysis, is required for validating our results, as well as for further exploring the predictive value of chitotriosidase levels on the long-term outcome in patients with cystinosis.

In conclusion, we demonstrate that plasma chitotriosidase enzyme activity has the potential to rule out poor disease control with a considerable predictive value and could therefore be used as a long-term therapeutic monitor and predictor of disease severity in nephropathic cystinosis.

Disclosures

Dr. Elmonem reports grants from ERA-Net, E-Rare2-JTC2014, outside the submitted work. Dr. Levtchenko reports grants from Horizon Pharma, during the conduct of the study; and other from Advincenne, Chiesi, Kyowa Kirin, and Recordati, outside the submitted work. All remaining authors have nothing to disclose.

Funding

The study was supported by Raptor Pharmaceuticals and Horizon Pharma from 2017 onwards. Dr. Veys is supported by the Research Foundation–Flanders (F.W.O. Vlaanderen) grant 11Y5216N. Dr. Elmonem is supported by ERA-Net, E-Rare2-JTC2014: Novel therapies for cystinosis. Dr. Prencipe is supported by Cystinosis Research Foundation grant CRFF-2016-006. Dr. Levtchenko is supported by F.W.O. Vlaanderen grant 1801110N, and the Cystinosis Research Network, Cystinosis Research Foundation, and Cystinosis Ireland.

Supplemental Material

This article contains the following supplemental material online at http://jasn.asnjournals.org/lookup/suppl/doi:10.1681/ASN.2019080774/-/DCSupplemental.

Supplemental Table 1. Defining criteria for the diagnosis of the five main established cystinosis ERCs.

Supplemental Table 2. Correlation analysis between the proposed biomarkers of macrophage activation and the number of ERCs in the subgroup of patients harboring at least one ERC.

Supplemental Figure 1. Median values for biomarkers of macrophage activation and WBC cystine values in ERC-positive patients in relation to the number of ERCs. Biomarker values ([A] IL-1β, [B] IL-6, [C] IL-18, [D] chitotriosidase enzyme activity) and WBC cystine levels were categorized according to the number of ERCs. P -values were on the basis of Kruskal–Wallis test. Data are presented as median±interquartile range.