Association of FKBP4 gene polymorphisms with glucocorticoid efficacy in pediatric systemic lupus erythematosus

Abstract

Background/Objectives

This study aimed to analyze the allele frequencies and genotype distributions of key FKBP4 single nucleotide polymorphisms (SNPs) (rs11833878, rs1981655, and rs41456246) in pediatric systemic lupus erythematosus (pSLE) patients versus healthy controls and to evaluate their association with glucocorticoid (GC) treatment efficacy.

Methods

Forty-five patients with pSLE and 45 age- and sex-matched healthy controls were recruited between October 2022 and June 2024. Genotyping was performed using the imLDR^® kit. The patients received standard GC treatment. Clinical assessments (SLEDAI-2K, laboratory parameters, and GC dosage) were performed at baseline, 3, and 6 months.

Results

Genotypes at rs11833878 and rs41456246 were significantly correlated with clinical indicators. rs11833878 was associated with a significant difference in baseline platelet (PLT) count (P = 0.037) and PLT improvement at 3 months (P = 0.016). From baseline to 6 months, it correlated with SLEDAI-2K reduction (P = 0.019) and PLT change (P = 0.035). Between 3 and 6 months, it was associated with a 24-hour urinary protein reduction (P = 0.047). rs41456246 was associated with a significant difference in baseline white blood cell (WBC) count (P = 0.011) and GC dosage at 3 months (P = 0.029). From baseline to 6 months, it correlated with the magnitude of GC tapering (P = 0.021) and PLT improvement (P = 0.034). Between 3 and 6 months, it was associated with the extent of GC tapering (P = 0.011). No significant associations were found for rs1981655.

Conclusion

FKBP4 polymorphisms may modulate the therapeutic response in pSLE, potentially influencing GC receptor function. rs11833878 was linked to improvements in PLT, SLEDAI-2K score, and proteinuria, whereas rs41456246 was associated with WBC, GC dosage, and PLT recovery. Genotyping these loci may serve as a valuable reference for developing personalized treatment strategies.

Introduction

Pediatric systemic lupus erythematosus (pSLE) is a multisystem autoimmune disease clinically characterized by the presence of various serum autoantibodies that can cause acute and chronic progressive damage to tissues and organs throughout the body [1]. Epidemiological data indicate that the prevalence of systemic lupus erythematosus (SLE) in the Chinese population ranges from 30 to 70 cases per 100,000 individuals. Of these, pSLE accounts for approximately 10%–20% of total SLE cases and constitutes 15%–25% of pediatric rheumatic diseases [2], suggesting that pSLE is a significant subtype of pediatric rheumatic disease in China. Compared with adult-onset patients, the disease course in children is often more severe, characterized by higher disease activity and a heavier medication burden. This underscores the need for more aggressive early intervention and individualized treatment strategies for pSLE [3–5]. More broadly, targeted and immunomodulatory treatment approaches have also been explored in other autoimmune diseases, such as rheumatoid arthritis, reflecting the growing interest in individualized therapeutic strategies for immune-mediated disorders [6].

Currently, major domestic and international clinical guidelines recommend glucocorticoids (GCs) as the first-line treatment for pSLE [7]. However, long-term use of GCs is significantly associated with various adverse drug events (ADEs), including metabolic disorders (e.g., osteoporosis and new-onset diabetes), cardiovascular complications (e.g., hypertension), and immunosuppression-related infections [8]. With the clinical introduction of biological agents such as belimumab, the concept of “steroid-free treatment” has gradually emerged [9, 10]. Nevertheless, clinical practice indicates that disease flares frequently occur in patients with pSLE during GC tapering, leading to the prevalent phenomenon of GC dependence. Consequently, current clinical strategies recommend maintenance therapy with the minimal effective dose for steroid-dependent patients to achieve an optimal balance between disease control and drug toxicity [10].

In recent years, genome-wide association studies (GWAS) and genome-wide linkage analyses have identified over 40 genes associated with SLE [11]. These include multiple single nucleotide polymorphism (SNP) loci significantly associated with the therapeutic response in pSLE, such as GR, NR3C1, ABCB1, FKBP5, MIF, HSP90AA1, and HSP90B1 genes [12–17]. These findings suggest that genetic factors play a pivotal regulatory role in the disease prognosis and therapeutic response of pSLE. As a crucial co-chaperone of HSP90 [18], FKBP4 participates in the functional maturation, ligand-binding capacity, and nuclear translocation of steroid hormone receptors (SHRs) through the formation of molecular chaperone complexes [19]. In-depth exploration of the molecular interaction networks of FKBP family proteins has revealed that FKBP4 and FKBP5 participate in maintaining multiple core physiological functions through synergistic regulatory mechanisms. These include the regulation of immune homeostasis, neuronal microtubule dynamics, SHR signal transduction, transmembrane protein homeostasis, and stress-adaptive responses. These molecular mechanisms provide a new molecular pathological basis for understanding the specific pathogenic roles of these two isoforms in autoimmune diseases, psychiatric disorders, tumorigenesis and reproductive endocrine disorders [20].

Numerous studies have demonstrated that polymorphisms in FKBP4 and FKBP5 are closely associated with therapeutic response and prognosis in autoimmune diseases. Although the role of FKBP4 in immune regulation has garnered significant attention, research on its polymorphisms, specifically in the context of pSLE, remains limited. Therefore, the present study aimed to investigate the distribution of key SNPs in the FKBP4 gene, including rs11833878, rs1981655, and rs41456246, among pSLE patients compared to healthy pediatric controls. Furthermore, this study aimed to analyze the association between these genetic variants and the efficacy of standard treatment regimens for pSLE. The findings are expected to provide novel genetic markers for personalized treatment approaches in pSLE and offer a theoretical basis for optimizing clinical strategies.

Materials and methods

Study population

Participants

This single-center, observational, prospective cohort study enrolled participants from the Department of Pediatrics, the First Affiliated Hospital of Xinjiang Medical University, between October 2022 and June 2024. A total of 45 children diagnosed with pSLE were enrolled in the SLE group, consisting of 11 males and 34 females, with a mean age of 10.98 ± 2.53 years. All pSLE cases fulfilled the 2019 European Alliance of Associations for Rheumatology (EULAR)/American College of Rheumatology (ACR) classification criteria [21]. For the control group, 45 age- and sex-matched healthy individuals were recruited from a physical examination center during the same period. The control group included 18 males and 27 females, with a mean age of 11.47 ± 2.25 years. There were no statistically significant differences between the two groups in terms of mean age (t = − 0.969, P = 0.335) or sex distribution (χ² = 2.493, P = 0.114). This study was approved by the Ethics Committee of the First Affiliated Hospital of Xinjiang Medical University (Approval No. 220722-05).

Inclusion criteria

Participants were eligible for inclusion if they met the following criteria: (1) age ≤ 16 years; (2) fulfillment of the 2019 European League Against Rheumatism/American College of Rheumatology classification criteria for SLE [22]; (3) newly diagnosed SLE at their first visit to our hospital; (4) no known history of allergy to glucocorticoid medications; and (5) complete clinical data available. Written informed consent was obtained from the parents and guardians of all participants.

Exclusion criteria

Patients were excluded from the study if they presented with (1) other rheumatic or autoimmune diseases, (2) severe systemic disorders such as diabetes mellitus or hyperthyroidism, (3) congenital disorders, (4) a history of organ transplantation, or (5) a requirement for organ function replacement therapy (e.g., regular dialysis) at the time of diagnosis.

Instruments and reagents

Key instruments used in this study included the 3730xl Genetic Analyzer and 2720 Thermal Cycler (Applied Biosystems, ABI, Foster City, CA, USA). Other laboratory equipment included a Gel Imaging System (Shanghai Peiqing Technology Co., Ltd., Shanghai, China), FR-110 UV Analysis Device (Shanghai Furi Technology Co., Ltd., Shanghai, China), centrifuge 5810R (Eppendorf, Hamburg, Germany), TD5A-WS Benchtop Low-Speed Centrifuge (Changsha Xiangyi Centrifuge Instrument Co., Ltd., Changsha, China), and Milli-Q Academic water purification system (Millipore, Billerica, MA, USA).

The primary reagents used for genotyping were the imLDR™ Genotyping Kit, Tag DNA ligase, and GeneScan™-600 LIZ™ Size Standard (GeneSky Biotechnologies Inc., Shanghai, China). PCR amplification was performed using PerfectStart^® Taq DNA Polymerase, 10× PerfectStart^® Taq buffer, and 10× GC Enhancer (TransGen Biotech, Beijing, China). Additional reagents included dNTP mix (Takara Biomedical Technology, Kusatsu, Japan), MgCl₂ (Sangon Biotech, Shanghai, China), and Hi-Di™ formamide (Applied Biosystems).

Sample collection and DNA extraction

Peripheral venous blood (2 mL) was collected from each participant into EDTA-anticoagulant tubes and stored at -80 °C until DNA extraction. Genomic DNA was isolated from whole blood using a magnetic bead-based separation method on a fully automated HF48 Nucleic Acid Purification System. Prior to automated extraction, the samples underwent a preprocessing step. Briefly, 600–700 µL of whole blood was mixed with 1 mL of Red Blood Cell (RBC) Lysis Buffer, vortexed at 100 rpm for 5 min, and centrifuged at 13,000 rpm for 1 min to discard the supernatant. After repeating the washing step, the remaining pellet was resuspended in 400 µL of RBC Lysis Buffer and 20 µL of Proteinase K (10 mg/mL). The suspension was then transferred to the HF48 instrument for automated extraction, according to the manufacturer’s standard protocol (Program C1001 or C102). The extracted genomic DNA was stored at -20 °C for subsequent analysis.

Genotyping

Primers were designed using the GeneMapper 4.1 software (Applied Biosystems, USA) (Table 1). Genotyping was performed using the improved multiplex ligation detection reaction (iMLDR^®) technique. The PCR reaction mixture (20 µL) consisted of 1× PerfectStart^® Taq buffer, 1× GC enhancer, 2.5 mM dNTPs, 25 mM MgCl₂, Hot Start Taq DNA polymerase, and 1 µL template DNA. Amplification was performed on a 2720 Thermal Cycler under the following conditions: initial denaturation at 95 °C for 2 min, followed by 11 cycles at 94 °C for 20 s, 65 °C for 40 s, and 72 °C for 1.5 min, and 24 cycles at 94 °C for 20 s, 59 °C for 30 s, and 72 °C for 1.5 min. PCR products were purified by incubation with Shrimp Alkaline Phosphatase (SAP) and Exonuclease I at 37 °C for 1 h, followed by enzyme inactivation at 75 °C for 15 min. The subsequent ligation reaction was performed at 94 °C for 1 min and 56 °C for 4 min for 38 cycles. Ligation products were analyzed on an ABI 3730XL Genetic Analyzer (Applied Biosystems, USA), and the data were processed using GeneMapper 4.1 software.

Table 1.

Primer sequences and ligation product lengths designed for FKBP4 genotyping

SNP	Polymorphism	Ligation Primer Sequence(5’-3’)	Ligation Product Length(bp)
rs11833878	G/C	RG: TCTCTCGGGTCAATTCGTCCTT CATTGTGGGAGCTCAGGGCATC RC: TGTTCGTGGGCCGGATTAGT CATTGTGGGAGCTCAGGGCATG RP: TAGCTCATCTGGCCTAGACTGGG TT	RG:90 RC:88
rs1981655	A/T	RA: TTCCGCGTTCGGACTGATAT CACAAGGAGTTGGGTAGAGTGGCTCT RT: TACGGTTATTCGGGCTCCTGT CACAAGGAGTTGGGTAGAGTGGCTCA RP: TTTTTTCAGTTCTTGGGAGCTGC	RA:91 RT:92
rs41456246	G/A	RG: TCTCTCGGGTCAATTCGTCCTT AGGGCAGCCAGGCATTGAGC RA: TGTTCGTGGGCCGGATTAGT AGGGCAGCCAGGCATTGAGT RC: TTCCGCGTTCGGACTGATAT AGGGCAGCCAGGCATTGAGG RP: TGAGATAAGGCCTGGCTCTTAGC TTTTTTTTTT	RG:94 RA:96 RC:94

Clinical data collection and assessment

Data collection and grouping

This study used a prospective cohort design. Clinical data were collected at three distinct time points: baseline (initiation of GC therapy), 3 months post-treatment, and 6 months post-treatment. The collected information included demographic characteristics, disease activity scores (SLEDAI-2K), daily glucocorticoid (prednisone) dosages, white blood cell (WBC) count, platelet (PLT) count, complement levels (C3 and C4), and 24-hour urinary protein quantification. Glucocorticoid dosage was recorded as the daily prednisone-equivalent dose in mg/day. When the original medical record documented dosage as the number of tablets, the value was converted to mg/day according to the tablet strength. In this cohort, one prednisone tablet contained 5 mg; therefore, the recorded tablet number was multiplied by 5 to obtain the prednisone-equivalent dose in mg/day. Steroid sensitivity was evaluated clinically and indirectly based on changes in prednisone-equivalent dose during follow-up together with changes in disease activity and laboratory parameters. For longitudinal comparisons, delta values were calculated as the earlier time-point value minus the later time-point value. For example, Δ baseline–month 3 was calculated as baseline value minus month 3 value, and Δ month 3–month 6 was calculated as month 3 value minus month 6 value. Therefore, negative Δ values indicate an increase over time, whereas positive Δ values indicate a decrease over time. For analysis, patients were stratified into distinct subgroups based on their FKBP4 genotypes to evaluate the association between genetic polymorphisms and the therapeutic efficacy of standard treatment regimens. Steroid-related adverse effects were not included as predefined study endpoints and were therefore not systematically collected or analyzed in this cohort.

Evaluation of therapeutic efficacy

Disease activity was evaluated using the SLEDAI-2K score, which was stratified as mild (≤ 6), moderate (7–12), or severe (> 12). The Lupus Low Disease Activity State (LLDAS) was defined by the simultaneous fulfillment of five criteria: (1) SLEDAI-2K score ≤ 4 with no activity in major organ systems or hemolytic anemia; (2) no new lupus disease activity; (3) a Physician Global Assessment (PGA) score ≤ 1; (4) a current prednisone dosage ≤ 7.5 mg/day; and (5) good tolerability of standard maintenance doses of immunosuppressive drugs.

Thrombocytopenia was defined as a platelet count < 100 × 10⁹/L. To avoid interpreting numerical platelet changes alone as therapeutic efficacy, additional categorical analyses were performed to evaluate the presence of thrombocytopenia at baseline and its persistence or resolution at 3 and 6 months. Resolution of thrombocytopenia was defined as recovery of platelet count to ≥ 100 × 10⁹/L during follow-up among patients who had thrombocytopenia at baseline. Because resistant thrombocytopenia was not predefined as a formal endpoint in the original study protocol, persistent thrombocytopenia during follow-up was analyzed descriptively rather than interpreted as confirmed treatment resistance.

Quality control

Strict quality control measures were implemented to ensure the integrity of the data. Selection bias was minimized by adhering to pre-defined inclusion and exclusion criteria. To address information bias, a multi-source verification mechanism was employed that integrated electronic medical records (EMR) with guardian interviews. Measurement bias was mitigated by using standardized data collection tools and published Pediatric SLE response/monitoring recommendations, including the PRINTO/ACR criteria for evaluating response to therapy in juvenile SLE and the SHARE recommendations for childhood-onset SLE [23, 24]. All personnel involved in data collection and analysis received uniform training on the standard operating procedures.

Statistical analysis

Data analysis was performed using IBM SPSS Statistics version 27.0. Continuous variables with a normal distribution are expressed as mean ± standard deviation (± s), and differences between groups were assessed using the Student’s t-test. Continuous variables with a non-normal distribution are presented as medians and interquartile ranges (IQR), and group comparisons were conducted using the Wilcoxon rank-sum test. Categorical variables were described using frequencies and percentages, and differences between groups were evaluated using Pearson’s χ² test or Fisher’s exact test as appropriate. The Hardy–Weinberg equilibrium (HWE) for genotype distributions was assessed using the χ² test. Statistical significance was set at p < 0.05. Because of the limited sample size and small genotype subgroups, multivariable regression models adjusting for treatment heterogeneity, systemic involvement, and autoantibody profiles were not performed; therefore, the reported associations were interpreted as exploratory and unadjusted.

Results

Hardy–Weinberg equilibrium and genotype distribution

Using the 1000 Genomes Project database and Haploview software (LD r² > 0.8, MAF > 0.01), six tag SNPs in the FKBP4 gene were initially identified: rs1981655, rs41456246, rs11833878, rs11062360, rs56196860, and rs73050592. However, three loci (rs11062360, rs56196860, and rs73050592) were excluded from the final association analysis due to extremely low counts of minor allele homozygotes (n = 0–2) in the study population, which did not meet the statistical requirements for valid genetic association studies [22, 23].

The remaining three loci (rs11833878, rs1981655, and rs41456246) were analyzed. The genotype distributions of these SNPs in both the SLE and control groups conformed to the Hardy–Weinberg equilibrium (P > 0.05). As shown in Tables 2, 3 and 4, there were no statistically significant differences in the genotype or allele frequency distributions of rs11833878, rs1981655, or rs41456246 between the SLE and healthy control groups (P > 0.05).

Table 2.

Distribution of allele and genotype frequencies of rs11833878 in SLE patients and healthy controls [n(%)]

Group	n	Genotype			Allele		HWE-P value
		GG	CG	CC	G	C
SLE group	45	1(2.2)	14(31.1)	30(66.7)	16(17.8)	74(82.2)	1.000
Control group	45	1(2.2)	14(31.1)	30(66.7)	16(17.8)	74(82.2)	1.000
X2 value /P-value		0.262/1.000			0.000/1.000

Table 3.

Distribution of allele and genotype frequencies of rs1981655 in SLE patients and healthy controls [n(%)]

Group	n	Genotype			Allele		HWE-P value
		TT	AT	AA	T	A
SLE group	45	2(4.4)	17(37.8)	26(57.8)	21(23.3)	69(76.7)	1.000
Control group	45	0(0.0)	17(37.8)	28(62.2)	17(18.9)	73(81.1)	0.320
X2 value/P-value		1.719/0.633			0.543/0.465

Table 4.

Distribution of allele and genotype frequencies of rs41456246 in SLE patients and healthy controls [n(%)]

Group	n	Genotype			Allele		HWE-P value
		AA	GA	GG	A	G
SLE group	5	0(0.0)	4(8.9)	41(91.1)	4(4.4)	86(95.6)	1.000
Control group	5	0(0.0)	8(17.8)	37(82.2)	8(8.9)	82(91.1)	1.000
X2 value/P-value		1.538/0.353			1.429/0.232

Association between FKBP4 polymorphisms and therapeutic efficacy

Following a 6-month standardized treatment regimen, 45 pediatric patients with pSLE completed full follow-up. To investigate potential pharmacogenetic differences, clinical parameters (WBC count, PLT count, complement C3/C4, 24-hour urinary protein, SLEDAI-2K scores, and GC dosage) were stratified according to the genotype of each locus. sBecause numerical changes in PLT count may not necessarily indicate clinically meaningful platelet response, we additionally assessed thrombocytopenia status descriptively using PLT < 100 × 10⁹/L as the threshold. Among patients with baseline thrombocytopenia, the number of patients showing resolution of thrombocytopenia by 3 and 6 months was reviewed across genotype groups. This assessment was used to complement the delta-based PLT analysis and to avoid interpreting platelet changes within the normal range as direct evidence of therapeutic efficacy.

rs11833878 and treatment response

Significant associations were observed between the rs11833878 genotype and several clinical indicators during follow-up (Table 5). Although rs11833878 was associated with a significant difference in baseline PLT count, with higher PLT counts observed in patients carrying the GG + GC genotype than in those carrying the CC genotype, this baseline difference was interpreted cautiously because pretreatment blood cell values may be influenced by disease severity and other clinical factors. During glucocorticoid treatment, patients with the CC genotype showed greater PLT improvement from baseline to 3 months and from baseline to 6 months. The CC genotype was also associated with a greater reduction in SLEDAI-2K score from baseline to 6 months. In addition, between 3 and 6 months, patients with the CC genotype showed a greater reduction in 24-hour urinary protein than those with the GG + GC genotype. Therefore, the post-treatment changes suggest that rs11833878 may be associated with genotype-specific patterns of hematological, disease-activity, and renal response during glucocorticoid therapy.

Table 5.

Comparison of relevant clinical/laboratory indicators among different genotypes of rs11833878 (Mean ± SD or Median [IQR])

Time	Parameter	GG + GC Genotype	CC Genotype	Z/t value	P value
Baseline	PLT count (109/L)	212.13 (149.00, 262.00)	157.30 (74.50, 195.75)	2.083	0.037
	SLEDAI score	18.60 (11.00, 35.00)	22.40 (14.00, 27.00)	1.194	0.233
Baseline vs. Month 3	Δ PLT count (109/L)	-58.47 (-117.00, 15.00)	-129.20 (-203.50, -92.00)	2.420	0.016
Baseline vs. Month 6	Δ SLEDAI score	13.73 ± 5.90	19.40 ± 7.99	-2.429	0.019
	Δ PLT count (109/L)	-64.80 (-162.00, -4.00)	-140.93 (-226.50, -91.50)	2.107	0.035
Month 3 vs. Month 6	Δ 24 h Urinary Protein	-0.03 (-0.03, 0.04)	0.60 (-0.01, 0.24)	1.988	0.047

Note: Only significant or relevant parameters are displayed for brevity. Δ values were calculated as the earlier time-point value minus the later time-point value; therefore, negative values indicate an increase over time and positive values indicate a decrease over time

rs41456246 and treatment response

For the rs41456246 locus, genotype-specific differences were observed in several treatment-related parameters (Table 6). Although rs41456246 was associated with a significant difference in baseline WBC count, with higher WBC counts observed in patients carrying the GA genotype than in those carrying the GG genotype, this baseline association was not interpreted as direct evidence of altered glucocorticoid response because pretreatment WBC counts may reflect baseline disease heterogeneity. After glucocorticoid treatment, patients with the GA genotype had a lower GC dosage at 3 months than those with the GG genotype. However, over the full 6-month treatment period, patients with the GG genotype showed greater GC dose reduction and greater PLT improvement than those with the GA genotype. Similarly, between 3 and 6 months, the GG genotype was associated with a greater extent of GC tapering. These post-treatment findings suggest that rs41456246 may be associated with genotype-specific patterns of GC tapering and platelet recovery.

Table 6.

Comparison of relevant clinical indicators among different rs41456246 genotypes (Mean ± SD or Median [IQR])

Time	Parameter	GA Genotype	GG Genotype	Z/t value	P value
Baseline	WBC count (109/L)	8.45 (6.32, 11.21)	4.32 (2.54, 4.76)	2.553	0.011
Month 3	GC Dosage (prednisone-equivalent mg/day)	2.88 ± 2.32	5.45 ± 2.17	-2.255	0.029
Baseline vs. Month 6	Δ GC Dosage (prednisone-equivalent mg/day)	1.50 ± 3.24	5.40 ± 3.11	-2.386	0.021
	Δ PLT count (109/L)	23.75 ± 244.40	-129.15 ± 120.62	2.194	0.034
Month 3 vs. Month 6	Δ GC Dosage (prednisone-equivalent mg/day)	-0.38 ± 3.35	2.45 ± 1.89	-2.662	0.011

Note: Δ indicates the change in value between specified time points (e.g., baseline vs. month 6). Data are presented as mean ± SD for normally distributed variables (e.g., GC Dosage) or median (IQR) for non-normally distributed variables (e.g., WBC count). Only statistically significant parameters (P > 0.05) are shown

rs1981655 and treatment response

In contrast to the other two loci, no statistically significant differences were observed for the rs1981655 polymorphism in relation to the absolute values or changes in WBC count, PLT count, complement levels, 24-hour urinary protein, or SLEDAI-2K scores at any time point (P > 0.05). The detailed data for this locus are presented in Table 7.

Table 7.

Comparison of relevant clinical indicators among different rs1981655 genotypes

Time	Parameter	TT + AT Genotype	AA Genotype	Z/t value	P value
Baseline	SLEDAI score	21.26 (13.00, 28.00)	21.04 (13.75, 26.25)	0.357	0.721
	PLT count (109/L)	182.74 (91.00, 241.00)	170.35 (74.50, 242.50)	0.368	0.713
Baseline vs. Month 6	Δ SLEDAI score	17.00 (11.00, 23.00)	17.88 (12.75, 20.25)	0.058	0.954
	Δ GC Dosage	5.37 ± 3.22	4.83 ± 3.37	0.542	0.591

Note: Δ values were calculated as the earlier time-point value minus the later time-point value; therefore, negative values indicate an increase over time and positive values indicate a decrease over time. Data are presented as median (IQR) or mean ± SD. No statistically significant differences (P < 0.05) were observed for any parameter at this locus; representative data are shown for comparison

Association between FKBP4 polymorphisms and low disease activity state (LLDAS)

At the 6-month follow-up, the study evaluated whether specific FKBP4 genotypes were associated with a higher likelihood of achieving a Lupus Low Disease Activity State (LLDAS). As shown in Table 8, there were no statistically significant differences in the proportion of patients who achieved LLDAS among the different genotypes of rs11833878 (χ² = 0.000, P = 1.000), rs1981655 (χ² = 0.116, P = 0.734), and rs41456246 (χ² = 0.952, P = 0.329). This suggests that while these polymorphisms influence specific clinical parameters (such as platelet counts or GC dosage), they may not independently determine the overall achievement of the composite LLDAS outcome in this cohort study.

Table 8.

Distribution of low disease activity state (LLDAS) achievement across FKBP4 genotypes after 6 months of treatment

Locus	Genotype	LLDAS Achieved [n (%)]		Test Statistics	P Value
		Yes	No
rs11833878	GG + GC	4 (4.3)	11 (10.7)	χ² = 0.000	1.000
	CC	9 (8.7)	21 (21.3)
rs1981655	TT + AT	6 (5.5)	13 (13.5)	χ² = 0.116	0.734
	AA	7 (7.5)	19 (18.5)
rs41456246	GA	2 (1.2)	2 (2.8)	χ² = 0.952	0.329
	GG	11 (11.8)	30 (29.2)

Data are presented as the number of patients (percentage of the total cohort). Differences were assessed using Pearson’s χ² or Fisher’s exact tests

Discussion

In this exploratory prospective cohort study, the analyzed FKBP4 polymorphisms were not significantly different between patients with pSLE and healthy controls, suggesting that these loci may not be strongly associated with pSLE susceptibility in this cohort. However, rs11833878 and rs41456246 showed genotype-specific associations with selected post-treatment clinical changes during glucocorticoids therapy, whereas rs1981655 showed no significant association with treatment-related outcomes. The main genotype-associated findings involved PLT recovery, SLEDAI-2K reduction, 24-hour urinary protein changes, and glucocorticoid dose tapering. Previous studies have suggested that glucocorticoid response may be influenced by genes involved in hormone metabolism, receptor-complex regulation, and inflammatory signaling pathways [25, 26].

These post-treatment associations are biologically plausible in view of the known involvement of FKBP4/FKBP52 in HSP90-associated glucocorticoid receptor regulation, receptor nuclear translocation, calcium-dependent signaling, and related immune/hematological cellular pathways [27]. In addition, FKBP family proteins have been implicated in protein folding, chaperone activity, receptor signaling, protein trafficking, transcriptional regulation, apoptosis, and steroid receptor-related pathways [28]. However, these references provide only mechanistic background. Because this study did not include functional experiments, the observed genotype-clinical associations should be interpreted as exploratory and hypothesis-generating rather than as evidence of a confirmed cellular mechanism.

In the present study, rs11833878 and rs41456246 were associated with significant differences in baseline PLT and WBC counts, respectively. These baseline hematological differences may reflect variation in the initial disease phenotype rather than direct differences in glucocorticoid therapeutic efficacy. Hematological abnormalities, including platelet and white blood cell changes, are recognized clinical features of SLE and are incorporated into disease classification and activity assessment systems [22, 29]. However, these pretreatment hematological findings should be interpreted cautiously because baseline blood cell values in pSLE may be affected by disease activity, inflammatory status, organ involvement, infection, and other clinical factors, rather than glucocorticoid receptor signaling alone. Therefore, the more relevant findings for the pharmacogenetic interpretation of FKBP4 are the genotype-associated changes observed after glucocorticoid treatment. In this regard, the CC genotype of rs11833878 was associated with greater PLT improvement, greater SLEDAI-2K reduction, and greater reduction in 24-hour urinary protein during follow-up, whereas the GG genotype of rs41456246 was associated with greater GC dose reduction and PLT improvement. These post-treatment associations are more consistent with the potential role of FKBP4 in glucocorticoid receptor signaling, HSP90-associated receptor regulation, and receptor nuclear translocation [27, 30, 31]. Nevertheless, because the sample size was limited and functional experiments were not performed, these findings should be considered exploratory and require validation in larger cohorts. Importantly, PLT changes were not interpreted as a standalone marker of therapeutic efficacy, because clinically meaningful platelet response depends on the presence and resolution of thrombocytopenia rather than numerical platelet change alone.

Through the analysis in this study, it was found that there was a statistically significant difference in the change in 24-hour urinary protein quantification between the 3-month and 6-month follow-up points of standard treatment among different genotype groups at the FKBP4 gene rs11833878 locus (P = 0.047 < 0.05). However, the difference in urinary protein quantification between baseline and the 6-month follow-up for this locus was not statistically significant, and no significant correlations were observed between other loci of this gene and urinary protein quantification. Previous studies have indicated that FKBP family members may play a crucial role in the formation of the glomerular filtration membrane. For instance, Yasuda et al. found that FKBP12 mRNA expression levels in the glomerulus are significantly higher than those in the renal cortex and medulla, and that FKBP12 localizes along F-actin within podocyte foot processes, participating in the formation of foot process structures and the slit diaphragm. As podocytes are a key component of the glomerular filtration barrier, FKBP12 may influence filtration membrane function by regulating the podocyte cytoskeleton. Given that FKBP4 and FKBP12 belong to the same FKBP family, we speculate that FKBP4 may possess similar functions and participate in the regulation of the glomerular filtration membrane (GFM). However, the specific expression pattern of FKBP4 in the kidney [32] and its impact on filtration membrane formation remain unclear. Although this study identified an association between the rs11833878 locus and short-term changes in urinary protein levels, the overall evidence remains insufficient to establish the definitive role of FKBP4 in renal pathophysiology. Therefore, future experiments, such as tissue localization analysis of FKBP4 and studies using loss-of-function or overexpression models, are needed to verify its potential mechanisms in glomerular filtration membrane formation and the pathogenesis of proteinuria.

The scientific value of this study warrants further validation across multiple dimensions. First, given the limitations of the single-center sample size (n = 45) and the longitudinal observation period (October 2022 to June 2024), it is essential to elevate the level of evidence through prospective multi-center cohort studies (with a target sample size of ≥ 300 cases and a follow-up duration of ≥ 3 years). Second, utilizing CRISPR/Cas9 gene-editing technology, which has been widely recognized as a potential tool for genome editing, could be used to construct humanized FKBP4 mutant mouse models for multi-level validation spanning molecular, cellular, and animal models, to evaluate the impact of rs11833878 allelic variations on GR signal transduction in the future [33, 34]. Finally, integrating single-cell multi-omics technologies with organoid culture systems could systematically elucidate the dose-response relationship between epigenetic modifications of FKBP4 (such as DNA methylation) and glucocorticoid sensitivity. These expanded research avenues will complete the evidence chain linking genetic polymorphisms to the clinical phenotypes.

The limitations of this study are as follows. The relatively small sample size and short observation period limit the representativeness of the results, introducing potential selection bias and risks regarding generalizability. As a prospective single-center study, center-specific selection bias may exist, and the results may be influenced by regional and population characteristics; thus, extrapolation of the findings requires caution. Treatment heterogeneity and differences in systemic involvement may also have confounded the observed changes in disease-activity and laboratory measures. Changes in PLT count, SLEDAI-2K score, urinary protein, and glucocorticoid tapering are not independent of organ involvement, concomitant therapies, or autoantibody profiles. In particular, antiphospholipid antibody status, including lupus anticoagulant, anticardiolipin antibodies, and anti-β2-glycoprotein I antibodies, may be relevant to thrombocytopenia and treatment resistance. Because of the limited sample size and small numbers within genotype subgroups, a fully adjusted multivariable regression model including treatment variables, systemic involvement, and autoantibody status would be underpowered and at risk of overfitting in this cohort. Therefore, the genotype–clinical associations observed here should be interpreted as unadjusted exploratory findings rather than independent effects of FKBP4 polymorphisms. The dimensions of efficacy data collection were limited, and the assessment indicators or follow-up duration may have been insufficient, potentially leading to an over-or underestimation of the therapeutic effect size. In addition, steroid-related adverse effects were not systematically evaluated because the study was primarily designed to assess genotype-associated differences in glucocorticoid efficacy rather than treatment safety; therefore, future studies should include standardized monitoring of glucocorticoid-related adverse events, including potential organ-related toxicity, as experimental evidence has shown that dexamethasone exposure can be used to model renal toxicity [35]. There is a lack of basic experimental evidence (e.g., molecular mechanisms and animal models) to corroborate the clinical conclusions; therefore, the biological plausibility and causal associations of the findings require further verification.

Conclusions

This study provides the first evidence that genetic polymorphisms in FKBP4 are significantly associated with the therapeutic efficacy of glucocorticoids in pediatric Systemic Lupus Erythematosus (pSLE). Specifically, we identified that rs11833878 and rs41456246 loci serve as potential biomarkers for predicting hematological recovery, renal improvement, and glucocorticoid sensitivity. Mechanistically, these associations may be driven by the role of FKBP4 in regulating the Glucocorticoid Receptor–HSP90 complex and Store-Operated Calcium Entry (SOCE) pathways, which are critical for platelet generation and immune regulation. Although the rs1981655 locus showed no correlation with treatment outcomes in this cohort, the positive findings for the other two loci highlight the value of FKBP4 genotyping in clinical practice. These markers could enable clinicians to stratify patients by risk, identify those requiring more aggressive monitoring for renal or hematological complications, and optimize glucocorticoid tapering schedules to minimize the toxicity. Future research should focus on validating these findings in multi-center cohorts and elucidating the precise molecular mechanisms through functional assays.

Author contributions

Conceptualization, H.Z.; methodology, X.W., N.Q.; software, X.W., N.Q.; validation, H.Z., L.Z.; formal analysis, X.W., L.Z.; investigation, X.W., L.Z., N.Q.; data curation, X.W., L.Z., N.Q.; writing—original draft preparation, X.W.; writing—review and editing, X.W., L.Z., N.Q.; supervision, H.Z., B.B.; project administration, H.Z., B.B.; funding acquisition, H.Z. All authors have read and agreed to the published version of this manuscript.

Funding

This research was funded by the General Program of the Natural Science Foundation of Xinjiang Uygur Autonomous Region (Project Number: 2022D01C466; Principal Investigator: Hongtao Zhu).

Data availability

The raw data supporting the conclusions of this article will be made available by the authors upon request.

Declarations

Institutional review board statement

The study was conducted in accordance with the Declaration of Helsinki and approved by the Medical Ethics Committee of the First Affiliated Hospital of Xinjiang Medical University (protocol code 220722-05 and date of approval: July 29, 2022).

Informed consent statement

Informed consent was obtained from all the participants involved in the study.

Competing interests

The authors declare no competing interests.