Comparative analysis of neutrophil gelatinase-associated lipocalin and other laboratory markers for lupus nephritis: a cross-sectional investigation

Abstract

Background

Lupus nephritis is one of the most serious complications of systemic lupus erythematosus. This study evaluates the prevalence and correlation between neutrophil gelatinase-associated lipocalin and other biomarkers associated with renal involvement in systemic lupus erythematosus.

Methods

Paired serum and urine specimens from 50 suspected systemic lupus erythematosus patients, characterized by antinuclear antibodies detected by indirect immunofluorescence assay and varying positive concentrations of anti-double stranded DNA antibodies by Crithidia luciliae immunofluorescence assay, were investigated. Of these 50 patients, 18 were identified with renal involvement based upon laboratory serology. Patients and healthy control serum samples (n = 50) were also evaluated for high avidity double stranded DNA IgG antibodies, anti-C1q IgG antibodies, and serum creatinine. The prevalence and relationship between biomarkers were evaluated using statistical methods.

Results

Serum and urine neutrophil gelatinase-associated lipocalin concentrations were significantly elevated in patients compared to controls, with a prevalence of 24% and 36%, respectively. These concentrations were also more markedly increased in systemic lupus erythematosus patients with renal involvement than those without. Spearman’s correlations between neutrophil gelatinase-associated lipocalin and other biomarkers tested ranged from 0.06 to 0.66 in all patients. Combined concordance as determined by Cronbach alpha coefficient between biomarkers was reduced from 0.71 to 0.58 (serum) and 0.62 (urine) when neutrophil gelatinase-associated lipocalin was removed.

Conclusions

Neutrophil gelatinase-associated lipocalin concentrations are elevated and demonstrate variable associated with other laboratory markers for renal involvement in systemic lupus erythematosus. Prospective longitudinal studies are needed to determine the optimal biomarker combinations for use in routine management of systemic lupus erythematosus patients at-risk for lupus nephritis.

Keywords

Neutrophil gelatinase-associated lipocalin systemic lupus erythematosus lupus nephritis biomarkers

Introduction

Systemic lupus erythematosus (SLE) is a chronic autoimmune disease characterized by heterogeneity in disease manifestations as well as diversity in immunologic and therapeutic responses. Despite longstanding research efforts, precise diagnosis, variable disease presentation and course, and predicting response to treatment remains problematic, due to the lack of sensitive and specific biomarkers.^1,2 A number of epigenetic, environmental, and hormonal factors have been reported to contribute to the immune abnormalities and clinical heterogeneity that occurs in SLE. It is likely that network interplay between these elements drives the production of a variety of autoantibodies, complement products, inflammatory markers, and other mediators identified as biomarkers to diagnose, monitor, stratify, and/or predict disease course, risk, or response to treatment.^1–7 Those of laboratory and clinical interest are diverse autoantibodies and complement products, which in addition to clinical manifestations such as skin lesions, arthritis, renal disorder, haematologic changes, and neurologic disorders, are traditionally considered hallmarks of the disease.^8,9

Of the different clinical manifestations associated with SLE, lupus nephritis (LN) is one of the most common and is associated with significant morbidity and mortality. In the US, approximately 35% of adults with SLE have clinical evidence of LN at the time of diagnosis, with an estimated 50–60% developing nephritis during the first 10 years of disease.¹⁰ Among these patients, the progression to end-stage renal disease is estimated at a rate of 5–10%.¹¹ Improved methods for detecting LN or renal involvement would allow earlier treatment, thereby preventing irreversible impairment of renal function and damage. In place of invasive, subjective, and costly serial renal biopsies, laboratory testings such as creatinine clearance, urine protein, complement factors 3 and 4 (C3, C4), serum creatinine, and anti-dsDNA titres have been used for decades to follow onset, course, severity and/or renal involvement. It is, however, recognized that these analyses are inadequate as they are limited in responsiveness to change and therefore unsuitable for optimal patient care and management.^6,10

In an effort to reliably predict LN, several candidate biomarkers have been identified.^1,2,6 Among the autoantibodies, antichromatin/antinucleosome, high avidity (HA) dsDNA and anti-C1q antibodies have shown some promise as biomarkers of disease activity and/or renal involvement. Of the urine protein biomarkers, neutrophil gelatinase-associated lipocalin (NGAL), soluble vascular cell adhesion molecule 1 (sVCAM-1), monocyte chemotactic protein 1 (MCP-1), and tumour necrosis factor-like weak inducer of apoptosis (TWEAK) amongst others have received considerable attention in predicting renal involvement in SLE.^1,2,6 Of these, NGAL, a small protein expressed in the neutrophils, certain epithelial cells, and the renal tubules in response to kidney injury, has been reported in a variety of acute and chronic diseases.^12–15 In healthy physiologic conditions, NGAL concentrations are minimally expressed in urine and plasma, but rapidly increase in response to kidney damage to reach diagnostic thresholds within a limited period of time. This is in contrast to the more routine SLE kidney biomarkers that assess renal function such as creatinine, where increased concentrations may not be observed until 24 to 48 h after injury and often lack both diagnostic sensitivity and specificity.

The objective of this study was to investigate the prevalence and relationship between NGAL and other laboratory biomarkers associated with renal involvement in a cohort of patients under evaluation for SLE in a reference laboratory setting.

Materials and methods

Patient population and heathy controls

Fifty patient samples, with paired serum and urine specimens, received at ARUP Laboratories for antinuclear antibody (ANA) by indirect immunofluorescence assay (IFA) and anti-dsDNA IgG antibody by IFA titre (Crithidia luciliae indirect fluorescent test, CLIFT; both from INOVA Diagnostics, San Diego, CA) were investigated. Samples from these patients with suspected SLE (sSLE) were selected based on a positive ANA by IFA with an associated homogenous pattern and/or a positive CLIFT result. Of the 50 patients (40 females, 10 males; 13–60 years of age) fitting the above criteria, 18 patients (16 females, 2 males; 19–60 years of age) were receiving treatment from the University of Utah Hospital, Salt Lake City, Utah. Following testing, a comprehensive, retrospective chart review was performed for these 18 patients to incorporate clinical SLE diagnosis (cSLE) and other laboratory evaluations for renal involvement. In this investigation, renal involvement was defined by abnormal protein concentration, red blood cells, or cellular casts in the urine analysis, and abnormal total protein-to-creatinine ratio in random urine collection at the time of lupus evaluation and/or an episode of biopsy-confirmed nephritis within a month of laboratory analysis. Results were interpreted as abnormal based on reference ranges provided by the laboratory for each analysis.

Paired serum and urine samples from 50 local, self-reported healthy volunteers (22–64 years of age; 33 female, 17 male) served as controls for this study. All studies involving the use of human specimens were approved by the University of Utah Institutional Review Board (Salt Lake City, UT; protocols 7275 & 29507).

Detection of serologic and urine biomarkers

Serum samples were tested for HA anti-dsDNA IgG antibodies (ELISA, INOVA Diagnostics), anti-C1q IgG antibodies (ELISA, INOVA Diagnostics), and enzymatic serum creatinine (MODULAR P, Roche Diagnostics, Indianapolis, IN). The presence and concentration of NGAL in serum and urine were detected using the NGAL Test reagent kit (BioPorto Diagnostics, Hellerup, Denmark), a particle-enhanced turbidimetric immunoassay with application parameters for several automated methods; for this study the Roche cobas c501 was used. All testings were performed at ARUP Laboratories following manufacturers’ instructions. Manufacturers’ recommended cut-off values were utilized for HA dsDNA (>30 IU/mL) and anti-C1q (≥40 units) assays. Reference ranges for serum creatinine (35.4–106.1 μmol/L), serum NGAL (51–265 ng/mL), and urine NGAL (2–113 ng/mL) were established at ARUP Laboratories.

Statistical analysis

The prevalence of the specific biomarkers in each group was established using manufacturers’ recommended or laboratory established cut-off values, as indicated. The D’Agostino-Pearson omnibus K2 normality test was performed using GraphPad Prism (version 5.04, GraphPad Software, Inc., La Jolla, CA) to determine if the datasets were normal. Significance, as determined by P-values <0.05 from the Mann–Whitney test, was calculated using GraphPad Prism. All other analyses including Spearman correlation coefficient and Cronbach alpha coefficient were determined using SAS (version 9.3, SAS Institute Inc., Cary, NC).

Results and discussion

Along with the identification of novel serologic and urine biomarkers for identifying and managing SLE patients at-risk for renal involvement, the examination of the relationships between established and emerging tests to determine optimal strategies for patient care is crucial. Current biomarkers have been reported to have limited clinical significance in early prediction of renal involvement in SLE.^1–3 In this preliminary cross-sectional laboratory investigation, the prevalence of serum and urine NGAL as well as routine tests for the evaluation of SLE was examined in a cohort of patients undergoing screening for SLE, along with a subset of patients with confirmed disease and renal involvement.

The prevalence of serum NGAL was 24% compared to 36% for urine NGAL in the sSLE group (Table 1). Thus, relative to the healthy control population, the positivity rates for both serum and urine NGAL in the unselected and confirmed SLE patients were significantly higher. It is important to note that the NGAL assay is robust in terms of repeatability. During internal evaluation studies performed at ARUP Laboratories, total CVs observed from a five-day period were ≤5.2% for samples ranging from 78.6 to 735.1 ng/mL (data not shown). For all biomarkers examined, the presence of HA dsDNA IgG antibodies was higher in the sSLE and cSLE groups (60% and 39%, respectively) than in the non-patient cohort. Of note, the prevalence of urine NGAL in the sSLE and cSLE groups was comparable to serum creatinine and anti-C1q antibodies. However, in the subset of confirmed SLE patients with renal involvement, only serum creatinine (40%) and urine NGAL (40%) showed comparable prevalence (Table 1). Furthermore, in the defined SLE cases as well as in those with renal involvement, the prevalence of serum NGAL was generally lower when compared to all patients and healthy controls.

Table 1.

Baseline characteristics of study population.

		High avidity dsDNA	Anti-C1q	Serum creatinine	Serum NGAL	Urine NGAL
	Cut-off	>30 IU/mL	≥40 Units	>106.1^a μmol/L	>265^a ng/mL	>113^a ng/mL
HCtrls	Median	1.0	4.0	70.7	90.0	17.0
(n = 50)	SD	2.4	19.2	17.7	56	35
	Range	0–14	2–97	53.0–114.9	28–401	1–211
	% positive	0	3	1	1	1
sSLE	Median	115.0	19.5	150.3	148.0	70.0
(n = 50)	SD	405.5	59.6	159.1	471.5	343.6
	Range	0–1508	3–215	44.2–848.6	6–2455	2–1900
	% positive	60	34	32	24	36
sSLE vs. HCrtls	P-value^b	<0.0001	<0.0001	0.0894	0.0005	0.0002
cSLE	Median	5.5	10.0	88.4	129.0	19.0
(n = 18)	SD	140.5	57.8	97.2	204.8	231.2
	Range	0–486	3–189	53.0–477.4	6–944	2–1001
	% positive	39	22	28	1	22
cSLE vs. HCrtls	P-value^b	0.1500	0.0205	0.0717	0.0849	0.2344
RI^c	Median	9.0	16.0	97.2	140.0	65.0
(n = 10)	SD	169.1	59.8	123.8	260.2	300.5
	Range	0–486	3–189	53.0–477.4	58–944	8–1001
	% positive	40	30	40	10	40
RI vs. HCrtls	P-value^b	0.2496	0.0994	0.1125	0.0105	0.0173

Cut-off values established at ARUP Laboratories.

P-values determined by the Mann–Whitney test.

Patients with RI came from the cSLE group.

HCtrls: Healthy controls; sSLE: suspected SLE; cSLE: confirmed SLE cases; RI: renal involvement; SD: standard deviations.

The demographic and laboratory features of the cSLE patients are outlined in Table 2. The majority of the cSLE population was female (92%) and Caucasian (67%). Due to the cross-sectional nature of this investigation, routine laboratory analyses for renal involvement were not consistently performed for all patients. The median antibody titres of the ANA by IFA and CLIFT were comparable in patients with and without renal involvement. As expected, the frequencies of patients with low C3 and/or C4 concentrations, random urine total protein-to-creatinine ratio and abnormal urine analysis were higher in those with renal involvement compared to those without.

Table 2.

Demographics and laboratory characteristics of patients with confirmed SLE (cSLE).

		All patients^a	With renal involvement	Without renal involvement
Demographic	Number of patients (n)	18	10	8
	Age in years, mean (SD)	37.3 (13.9)	38.0 (14.4)	36.5 (14.1)
	Gender, female	16/18 (92%)	10/10 (100%)	6/8 (75%)
	Caucasian	12/18 (67%)	6/10 (60%)	6/8 (75%)
	Non-Caucasian	6/18 (33%)	4/10 (40%)	2/8 (25%)
Routine laboratory evaluation	Patients with ANA+ results (%)	16 (88.9%)	9 (90%)	7 (87.5%)
	ANA by IFA, median titre (range)	1:320 (<1:40–1:10240)	1:200 (<1:40–1:10240)	1:320 (<1:40–1:5120)
	Patients with CLIFT+ (%)	18 (100%)	10 (100%)	8 (100%)
	CLIFT, median titre (range)	1:60 (1:10–1:2560)	1:60 (1:10–1:320)	1:60 (1:10–1:2560)
	Low C3 (<88 mg/dL)	4/13 (31%)	4/9 (44%)	0/4 (0%)
	Low C4 (<10 mg/dL)	3/13 (23%)	3/9 (33%)	0/4 (0%)
	RDW, mean (range)	14.2 (12.2–19.0)	14.7 (12.2–19.0)	13.5 (12.8–14.5)
	Random urine total protein-to-creatinine ratio, mg/g, mean (range)^b	1405 (43–5343)	1990 (59–5343)	89 (43–139)
	Urine analysis
	>5 RBC per Hpf	1/18 (50%)	1/10 (10%)	0/8 (0%)
	>5 WBC per Hpf	4/18(22.2%)	4/10 (40%)	0/8 (0%)
	Urine cast, positive	7/18 (38.9%)	7/10 (70%)	0/8 (0%)
	Abnormal urine protein	8/18 (44.4%)	8/10 (80%)	0/8 (%)
Biomarkers evaluation	HA dsDNA, mean (range)	88.1 (0–486)	111.4 (0–486)	58.9 (0–275)
	Anti-C1Q, mean (range)	35.9 (3–189)	39.5 (3–189)	31.5 (3–176)
	Serum creatinine, mean (range)	114.9 (53.0–477.4)	132.6 (53.0–477.4)	88.4 (53.0–477.4)
	Serum NGAL, mean (range)	165.7 (6–944)	217.4 (58–944)	101.0 (6–239)
	Urine NGAL, mean (range)	101.7 (2–1001)	164.4 (8–1001)	23.3 (2–103)

SD: standard deviation; C3: complement 3; C4: complement 4; RDW: red blood cell distribution width; Hpf: high power field.

Note: No significant differences were observed for any analyte between patients medians with or without renal involvement (P-value >0.05), except for urine protein (P-value 0.04).

All results for routine laboratory evaluation were not available for some patients. The total n was changed to account for this in order to calculate percentages.

Random urine total protein-to-creatinine ratio reference range: male 15–68 mg/g; female: 10–107 mg/g (18 years and older).

Of the five biomarkers tested, all except the anti-C1Q IgG test demonstrated significantly higher concentrations in patients with renal involvement compared to those without (Table 2). This observation may be limited by the cross-sectional nature of this investigation, the small sample size of the cohort as well as the lack of measures to define active LN and global disease activity. In a recent study by Bock et al.,¹⁶ anti-C1Q antibodies correlated not only with urine protein-to-creatinine ratio in patients with confirmed but also in those without confirmed renal involvement suggesting subclinical renal involvement. The observation that dsDNA IgG, serum creatinine, or urine NGAL concentrations are significantly elevated in patients with renal involvement than those without appears to be consistent with previous studies.^13–21 Of the patients with renal involvement, only two patients were positive for a single biomarker (urine NGAL or HA dsDNA) of the five evaluated. All other patients tested positive for at least two or more biomarkers with urine NGAL and serum creatinine or HA dsDNA and anti-C1Q co-occurring in 60% of the cases.

To define the relationships between biomarkers, any two pair was evaluated for degree of association by Spearman correlation coefficient, rho (Table 3). Three significant associations were noted, the most notable between HA dsDNA IgG and anti-C1Q IgG antibodies (P-value 0.0002). As expected, both urine and serum NGAL also showed signification correlation as demonstrated by rho of 0.589. Compared to urine NGAL, the association between serum NGAL and serum creatinine was more remarkable (0.442 versus 0.265 for serum and urine NGAL, respectively). Indeed, in patients with renal involvement, the Spearman’s correlation between serum NGAL and serum creatinine was quite substantial (0.78, data not shown). Nevertheless, of the all biomarkers, serum NGAL appears to be the least sensitive. The inter-relatedness between the biomarkers in the sSLE cohort was determined using Cronbach alpha coefficient. The estimated alpha for this cohort was 0.71 which demonstrates acceptable but not absolute correlation.¹⁷ Indeed, this value was reduced when NGAL was removed from the combined concordance analysis (0.58 and 0.62 for serum and urine NGAL, respectively). These results indicate that neither serum nor urine NGAL alone are sufficient in assessing SLE patients for renal involvement, but rather, could be useful when paired with other biomarkers. Therefore, this observation is strongly in support of the notion that a combination of biomarkers and not one may be sufficient to reliably diagnose renal involvement.

Table 3.

Correlation between NGAL and other SLE biomarkers investigated.

Biomarkers	Anti-C1Q IgG	Serum creatinine	Serum NGAL	Urine NGAL
HA dsDNA	0.660^a	0.06	0.213	0.254
Anti-C1Q IgG		0.134	0.061	0.195
Serum creatinine			0.442^a	0.265
Serum NGAL				0.589^a

NGAL: neutrophil gelatinase-associated lipocalin.

Note: Comparisons between NGAL and indicated biomarkers were estimated by Spearman correlation coefficient in the suspected patient group (n = 50) and clinically defined cases with renal involvement. A value of 1.0 represents ideal correlation.

P-value <0.05

With the recent developments in biomarker discovery for predicting renal involvement in SLE, a number of investigations have examined the associations between specific urinary biomarkers as well as serum protein profiles and histological features of nephritis in an attempt to reliably predict disease.^{3,7,14,15,18–20} One of the seminal studies involving paediatric SLE patients, a population that is most adversely affected by LN, was reported by Brunner and colleagues.²¹ Based on the report from this investigation, the authors concluded that presence of defined urine proteins could predict specific LN signatures such as activity and chronicity. Although the current investigation is limited by the low number of cSLE patients with renal involvement, ethnic diversity, and disease activity score, there was a general trend for patients with renal involvement as defined by current laboratory tools to have higher levels and/or frequencies of all, if not most, biomarkers associated with LN. As previously reported by others and shown here, no single biomarker is sufficient in identifying and/or stratifying SLE patients at-risk for renal involvement.^3,7,19 While SLE disease heterogeneity may partly be responsible for these observations, the prevalence of individual biomarkers lacks optimal sensitivity.

In conclusion, this study demonstrates that urine NGAL determination in addition to currently available laboratory tests for the evaluation of SLE may be useful in assessing renal involvement. Biomarker diversity and disease heterogeneity demands an informed approach geared towards the validation of biomarker ‘panels’ of renal involvement in SLE. Studies with well-defined endpoints that ensure a true measure of the intended clinical process in diverse cohorts coupled with robust analytical assays and high statistical power to confirm these panels are needed.

Footnotes

Acknowledgements

The authors thank Elisabeth Malmberg and Andrew Wilson for assistance with statistical analysis and Joshua Hunsaker for sample procurement and de-identification. The authors also acknowledge the provision of reagents from BioPorto Diagnostics and INOVA Diagnostics.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

Support for this study was provided by the ARUP Institute for Clinical and Experimental Pathology.

Ethical approval

The University of Utah Institutional Review Board approved this study (ref# 7275 & 29507).

Guarantor

AET.

Contributorship

SLL wrote the first draft of the manuscript, researched literature, performed testing, and data analysis. BBS was involved in protocol development, sample acquisition, and testing. KWD performed testing. JAS provided expertise in protocol development and data analysis. AET conceived the study, gained ethical approval, researched literature, performed data analysis and review. All authors reviewed and edited the manuscript and approved the final version of the manuscript.

References

Illei

Tackey

Lapteva

. Biomarkers in systemic lupus erythematosus: II. Markers of disease activity. Arthritis Rheum 2004; 50: 2048–2065.

Liu

Kao

Manzi

. Biomarkers in systemic lupus erythematosus: challenges and prospects for the future. Ther Adv Musculoskelet Dis 2013; 5: 210–233.

Cozzani

Drosera

Gasparini

. Serology of lupus erythematosus: correlation between immunopathological features and clinical aspects. Autoimmune Dis 2014; 2014: 321359–321359.

Jeltsch-David

Muller

. Neuropsychiatric systemic lupus erythematosus: pathogenesis and biomarkers. Nat Rev Neurol 2014; 10: 579–596.

Leffler

Bengtsson

Blom

. The complement system in systemic lupus erythematosus: an update. Ann Rheumat Dis 2014; 73: 1601–1606.

Parodi

Bennett

Brunner

. Biomarkers and updates on pediatrics lupus nephritis. Autoimmune Dis 2013; 39: 833–853.

Petri

. Is antibody clustering predictive of clinical subsets and damage in systemic lupus erythematosus? Arthritis Rheum 2005; 52: 4003–4010.

Hochberg

. Updating the American College of Rheumatology revised criteria for the classification of systemic lupus erythematosus. Arthritis Rheum 1997; 40: 1725–1725.

Petri

Orbai

Alarcon

. Derivation and validation of the systemic lupus international collaborating clinics classification criteria for systemic lupus erythematosus. Arthritis Rheum 2012; 64: 2677–2686.

10.

Hahn

McMahon

Wilkinson

. American College of Rheumatology guidelines for screening, treatment, and management of lupus nephritis. Arthritis Care Res 2012; 64: 797–808.

11.

Fiehn

. Early diagnosis and treatment in lupus nephritis: how we can influence the risk for terminal renal failure. J Rheumatol 2006; 33: 1464–1466.

12.

Bolignano

Donato

Coppolino

. Neutrophil gelatinase-associated lipocalin (NGAL) as a marker of kidney damage. Am J Kidney Dis 2008; 52: 595–605.

13.

Haase

Haase-Fielitz

Bellomo

. Neutrophil gelatinase-associated lipocalin as a marker of acute renal disease. Curr Opin Hematol 2011; 18: 11–18.

14.

Hinze

Suzuki

Klein-Gitelman

. Neutrophil gelatinase-associated lipocalin is a predictor of the course of global and renal childhood-onset systemic lupus erythematosus disease activity. Arthritis Rheum 2009; 60: 2772–2781.

15.

Rubinstein

Pitashny

Levine

. Urinary neutrophil gelatinase-associated lipocalin as a novel biomarker for disease activity in lupus nephritis. Rheumatology 2010; 49: 960–971.

16.

Bock

Heijnen

Trendelenburg

. Anti-C1q antibodies as a follow-up marker in SLE patients. PloS One 2015; 10: e0123572–e0123572.

17.

Tavakol

MDR

. Making sense of Cronbach's alpha. Int J Med Educ 2011; 2: 53–55.

18.

Alharazy

Kong

Mohd

. The role of urinary neutrophil gelatinase-associated lipocalin in lupus nephritis. Clin Chim Acta 2013; 425: 163–168.

19.

Brunner

Bennett

Mina

. Association of noninvasively measured renal protein biomarkers with histologic features of lupus nephritis. Arthritis Rheum 2012; 64: 2687–2697.

20.

Pitashny

Schwartz

Qing

. Urinary lipocalin-2 is associated with renal disease activity in human lupus nephritis. Arthritis Rheum 2007; 56: 1894–1903.

21.

Brunner HI, Mueller M, Rutherford C, et al. Urinary neutrophil gelatinase-associated lipocalin as a biomarker of nephritis in childhood-onset systemic lupus erythematosus. Arthritis Rheum 2006; 54: 2577–2584.