Introduction

Lupus nephritis (LN) is the most common severe manifestation of systemic lupus erythematosus (SLE). Despite the management of corticosteroids and aggressive immunosuppressive agents, a substantial percentage of LN patients still progress to chronic kidney disease (CKD), even end-stage renal disease (ESRD) [1]. Patients with diffuse proliferative LN, including pure PLN (class III or class IV) and mixed PLN (class III, IV mixed with class V), are at a high risk of progressing to CKD. Therefore, proper management of proliferative LN is primarily important in the control and prevention of ESRD in LN patients [2].

The severity of tubulointerstitial inflammation is known to be associated with poor renal prognosis [3, 4]. The number of interstitial infiltrates is one of the typical features used to evaluate tubulointerstitial inflammation. According to the modified histologic criteria of LN established by the International Society of Nephrology and the Renal Pathology Society (ISN/RPS), interstitial infiltrates are also one of the histological activity scoring indexes in LN [5]. Therefore, determination of the clinicopathological significance of interstitial infiltrates has great value.

Macrophage, T cell, B cell, and dendritic cell constitute the major compartments of interstitial infiltrates in the kidneys of LN, among which macrophage has the largest proportion, accounting for about 50% [6]. Macrophage plays a leading role in both renal injury and repair. During acute inflammation, macrophages first recruit from the peripheral blood, then exhibit a proinflammatory phenotype and finally switch to a reparative phenotype [7]. As shown in a NZB/NZW lupus mouse model, residential macrophage also contributed to renal fibrosis [8]. Increased renal macrophage infiltration was found to be associated with poor renal survival in a MRL-lpr lupus mouse model [9]. Dias et al. reported that renal CD68 macrophage was associated with progression to renal function decline in human proliferative LN, and interstitial expression of CD68 was positively correlated with SCr during the follow-up period [10]. However, the relationship between the histological lesions and the clinicopathological features of LN patients and CD68 macrophage has not been defined clearly.

The aim of this study was to examine the correlation between CD68 macrophage obtained from in LN patients by tubulointerstitium biopsy and the clinicopathological parameters and investigate the association between CD68 macrophage and prognosis of LN patients in a single center in China from 2016 to 2018.

Materials and methods

Patient selection

Enrolled in this retrospective study were 76 proliferative LN patients (class III, IV, III + V, or IV + V) classified according to the ISN/RPS classification criteria in the Division of Nephrology, Kidney Institute of CPLA, Changzheng Hospital Kidney Institute of CPLA and Division of Nephrology, Changzheng Hospital (Shanghai, China) from January 2016 to December 2018. Patients with overlapping Sjogren’s syndrome, vasculitis, rheumatoid arthritis, or other diffuse connective tissue diseases and those with biopsy-confirmed diabetic or hypertensive nephropathy were excluded from the study. The study protocol was approved by the Ethics Committee of Changzheng Hospital. Written consent was not required for this noninvasive retrospective study, as determined by the ethics committee.

Data collection

Patient data were collected from the medical records, including age, sex, and disease duration. The laboratory data included blood routine, 24-h urine protein, serum creatinine (SCr), complement 3 (C3), and the glomerular filtration rate (GFR) estimated by the Modification of Diet in Renal Disease (MDRD) equation. Autoantibodies were measured, including anti-double-stranded (ds) DNA, anti-Smith (Sm), anti-SSA/Ro, anti-SSB/La, anti-ribonucleoprotein (RNP), and antiphospholipid antibodies. All the biochemical data were collected within 3 days before renal biopsy. SLE disease activity index (SLEDAI) [11] of all patients was assessed at the time of admission to hospital.

Renal pathology

All patients were subjected to renal biopsy for light microscopy and immunofluorescence (IF). Only samples containing 10 or more glomeruli were considered acceptable for analysis. The following indices were assessed by light microscopy: the percentage of overall and segmental sclerosis and crescents, tubulointerstitial inflammation, and the extent of interstitial fibrosis/tubular atrophy (IFTA). IFTA was graded on a semiquantitative scale of 0 (none), 1–25% (mild), 26–50% (moderate), or > 50% (severe). The activity index (AI) and chronicity index (CI) were calculated based on the modified NIH system [5]. CD68 macrophages were counted per mm2 after anti-CD68 monoclonal antibody (Dako) staining. CD68 + cells were counted in the tubulointerstitial compartments by a light microscope at × 400 magnification. Counting cells per mm2 of stained tissues were evaluated by computer-assisted image analysis software (Count, Biomas, Erlangen, Germany) or by using a grid. All the above assessments were performed independently by two nephrologists.

Renal outcomes

Patients were divided into a renal progression group and a non-progression group. Renal progression was defined as a two-fold increase in SCr during the follow up, initiation of dialysis, transplantation, or death.

Statistical analysis

Clinical variables are presented as numbers and relative frequencies or medians and interquartile ranges. The unpaired t test or the Mann–Whitney test was used to analyze continuous data, as appropriate and categorical data using the chi-square test. Spearman’s correlation analysis was used to measure correlation between two parameters. All factors with P < 0.1 in univariate analysis were selected for Cox multivariate forward stepwise analysis. A two-tailed p value < 0.05 was considered statistically significant. All statistical analyses were performed using SPSS 25 software (IBM Corp., Armonk, NY, USA).

Results

Baseline clinical characteristics, laboratory and pathological findings, and treatment

Included in this study were 76 patients with biopsy-proven proliferative LN, including 11 with class III, 32 with class IV, 13 with class III + V, and 19 with class IV + V. The mean age was 32.6 ± 11.5, with the female gender predominating (90.7%). The mean duration of SLE and renal disease was 35.5 ± 53.4 and 22.1 ± 36.5, respectively. Hypertension was diagnosed in six patients, and no patient was diagnosed with diabetes. The median SCr level, proteinuria, and C3 level at renal biopsy was 0.76 mg/dL, 2.25 g/24 h, and 0.415 g/L, respectively. GFR < 60 mL/min/1.73 m2 at the time of the biopsy was detected in seven patients (9.3%). The anti-ds-DNA and anti-Smith antibody were positive in and 47 and 26 patients. ACL antibody was detected in seven patients.

The mean number of infiltrating CD68 macrophages was 381.7 ± 263.2/mm2, showing no significant difference in CD68 macrophages between these classes of LN. The median glomeruli sclerosis and crescents were 5.35% and 5%. The mild, modest, and severe IFTA were found in 24, 4, and 1 patient, respectively. Arteriolar sclerosis was detected in 34 patients. Notably, thrombotic microangiopathy was found in three patients. Overall, the mean AI and CI were evaluated by modified NIH activity, and chronicity indices were 6.78 ± 3.27 and 2.35 ± 1.71, respectively.

Individualized induction therapy was given according to the severity of systemic inflammation and histological lesions. Briefly, 53 patients received multiphase impulse treatment. Prednisone combined with mycophenolate mofetil (MMF) was given in 22 cases, multi-target therapy (MMF combined with calcineurin inhibitors) in 35 cases, cyclophosphamide (CTX) in 15 cases, and rituximab in three cases. During the maintenance period, prednisone was tapered gradually, combined with the immunosuppressive drug used in the induction period. HCQ and renin-angiotensin system blockers were administered in all patients with no contraindication. The detailed clinical characteristics, laboratory test, and histological findings and treatment regimens are shown in Table 1.

Table 1 Baseline characteristics of LN patients in each class
Full size table

Correlations between tubulointerstitial CD68 macrophage and clinicopathological features

The correlation between tubulointerstitial CD68 macrophages and clinicopathological features was analyzed by Spearman’s correlation analysis. The result showed that the number of tubulointerstitial CD68 macrophages was positively correlated with SCr level and negatively correlated with GFR (Fig. 1A). In addition, the number of tubulointerstitial CD68 macrophages was positively correlated with the rate of glomeruli sclerosis, focal segmental sclerosis, TI, and chronicity index (Fig. 1B). However, no significant correlation was observed between CD68 macrophages and SLEDAI and AI (Fig. 1C).

Fig. 1
figure 1

The correlation between clinicopathological parameters of 76 proliferative LN patients and the number of tubulointerstitial CD68 macrophages: A the number of tubulointerstitial CD68 macrophages positive correlated with the level of SCr, negative with GFR; B the number of tubulointerstitial CD68 macrophages positively correlated with pathological lesions including glomeruli sclerosis, focal segmental sclerosis, TI, and CI; and C no significant correlation between the number of tubulointerstitial CD68 macrophages and SLEDAI and AI. SCr, serum creatinine; GFR, glomerular filtration rate; TI, tubulointerstitial inflammation; CI, chronicity index; AI, activity index; and SLEDAI, SLE disease activity index

Full size image

Renal outcomes

Of the eight patients in ESRD progression group, five (6.6%) progressed to dialysis, three (3.9%) had a twofold increase in SCr, and no patient died during a mean follow-up period of 45 months. The results of comparison in the clinicopathological features between ESRD progression and non-ESRD progression groups are shown in Table 2. The level of CD68 macrophage infiltration in patients of ESRD progression group was significantly higher than that in non-ESRD progression group. Besides the SCr level and the degree of CI, TI in ESRD progression group was significantly higher, and the GFR was significantly lower in ESRD progression group as compared with non-ERSD progression group.

Table 2 Comparisons of clinicopathological features of LN patients in ERSD progression and non-progression groups
Full size table

Higher number of tubulointerstitial CD68 macrophages is associated with poor outcomes

Six statistically significant variables in univariate Cox regression analysis were subjected to multivariate Cox regression. They included SCr (HR = 1.255, P = 0.062), IFTA (HR = 0.015, P = 1.048), TI (HR = 1.05, P = 0.014), GFR (HR = 0.982, P = 0.044), CI (HR = 1.39, P = 0.039), and the number of tubulointerstitial CD68 macrophages (HR = 1.002, P = 0.011). The result of multivariate Cox regression analysis showed that only the number of tubulointerstitial CD68 macrophages was an independent variable associated with poor outcomes of LN (HR = 1.002, P = 0.012) (Table 3).

Table 3 Factors associated with poor renal outcome in LN patients as determined by Cox univariate and multivariate analyses (model 1)
Full size table

The cutoff value of tubulointerstitial CD68 macrophages was 340/mm2 in our study. The AUC of ESRD progression was 0.75, offering an 87.5% sensitivity and a 61.8% specificity for ESRD prediction within 4 years (Fig. 2). Based on the cutoff value of tubulointerstitial CD68 macrophages, univariate and multivariate Cox regression analyses were performed again. The result of univariate Cox regression analysis showed that tubulointerstitial CD68 macrophage > 340/mm2 was associated with ESRD progression (HR = 6.816, P = 0.078). However, only IFTA (HR = 0.014, P = 0.014) was independently associated with ESRD progression (Table 4).

Fig. 2
figure 2

ROC curve analysis. ROC, receiver operating characteristic; AUC, area under the curve; CI, confidence interval

Full size image
Table 4 Factors associated with poor renal outcome in LN patients as determined by Cox univariate and multivariate analyses (model 2)
Full size table

Discussion

Macrophage is known to be involved in renal damage of LN [12]. Our study showed that the number of tubulointerstitial CD68 macrophages was positively correlated with the active renal damage process, such as the degree of TI, but also with chronic progress including IFTA and CI. In addition, we showed that the higher number of tubulointerstitial CD68 macrophage was associated with ESRD progression in LN.

Previous studies have demonstrated that macrophages constitute most infiltrating immune cells [6, 13, 14]. A recent study by Kim et al. showed that CD11c + macrophages in the tubulointerstitium and urine of LN patients were abundantly present [15]. They suggested that these urinary CD11c + macrophages may play a pathological role in LN in that they were immunologically active and interacted with tubular epithelial cells [16]. In addition, urine CD11c + macrophages correlated with chronicity indices may be associated with the response of induction treatment with immunosuppressants [15]. We used the pan-macrophage marker CD68 to explore the potential pathological role in LN and found that these macrophages were correlated not only with chronic lesions but with TI.

Renal residential macrophages are highly pleiotropic cells and can rapidly be polarized in vitro to multiple phenotypes in response to the renal microenvironment [17, 18]. M1 macrophages produce inflammatory cytokines and interact with local DAMPs secreted by tubular epithelial cells during acute phase, while M2 macrophages are responsible for repair and regeneration of renal tissues [19, 20]. The association of renal macrophages with poor disease outcomes and renal fibrosis has been reported in both human renal diseases and mouse renal injury models. However, the full picture of pathogenic and protective roles of macrophages in LN and how they interact with local cells remain incompletely understood due to the plasticity of macrophages [21,22,23]. Future studies are required to further identify subgroups of macrophages with specific phenotypes mediated in the pathogenesis of proliferative LN.

Our study demonstrated that tubulointerstitial CD68 macrophages were positively correlated with baseline SCr, which is consistent with the study of Dias et al. [10]. We further suggested that when LN patients presented with acute renal injury, infiltrating immune cells including CD68 macrophages in the tubulointerstitium should be considered as one of the main reasons [23]. As these cells are highly involved in the development of chronic renal injury, urgent medications such as glucocorticoid pulse therapy may be necessary, especially when renal biopsy are not available.

Moderate/severe IFTA and vascular injury are known as poor renal prognostic factors associated with proliferative LN [3, 24]. Dias et al. did not show the prognostic significance of renal CD68 macrophages in proliferative LN [10]. Our Cox regression analysis results firstly suggested that the number of tubulointerstitial CD68 macrophages was independently associated with poor renal outcomes. In addition, the risk of progression to ESRD was high in LN patients with tubulointerstitial CD68 macrophage > 340/mm2. For this reason, special monitoring such as urgent immunosuppression should be given. In their 197 LN patients, Zhang et al. demonstrated that CD8 + T cell infiltrates were an independent poor prognostic factor, which further suggests that infiltrating immune cells in renal of LN need to be comprehensively studied [6]. However, we did not find vascular injury as a poor factor probably because only three patients were diagnosed with TMA in our series.

Our study had several limitations. Firstly, the number of patients in ESRD progression group was relatively small, and, therefore, some statistical results need to be interpreted with caution. Secondly, we did not perform tubulointerstitial CD68 macrophage subgroup and phenotype analyses on how these cells were mediated in LN. Finally, we were unable to obtain complete information about the effect of cumulative use of glucocorticoids and medical compliance on the part of the patients, although there was no significant difference in the use of medications between the two groups.

In conclusion, tubulointerstitial CD68 macrophage was a strong indicator of ESRD progression in proliferative LN. Early interventions are suggested when renal biopsy shows the presence of abundant tubulointerstitial CD68 macrophages.