Tubulointerstitial nephritis in primary Sjögren syndrome: clinical manifestations and response to treatment

Patients were referred to the UCL Centre for Nephrology tubular clinic and all underwent investigation with serum biochemistry, immunology, and renal biopsy. Urinary acidification testing was performed in 9 patients. GFR was estimated in all patients using the Modification of Renal Diet (MDRD eGFR) equation. 6 patients underwent 3-dose ⁵¹Chromium EDTA-GFR (⁵¹Cr-GFR) measurements before and after treatment. All patients with pSS who underwent renal biopsy from January 2007-December 2014 were included in this analysis.

The diagnosis of pSS was based on the 2002 American European consensus criteria [7], all patients satisfied the same criteria; ocular and oral symptoms, positive Schirmer test and positive Ro/La antibody status.

All patients who were suspected of having distal renal tubular acidosis (dRTA) underwent urinary acidification testing with a furosemide and fludrocortisone test [8] or an ammonium chloride test [9]. Briefly, the patient is administered either 40 mg of furosemide and 1 mg of fludrocortisone, or 0.1 mg/kg of ammonium chloride orally and the urine pH is monitored hourly. A fall in the urine pH to less than 5.3 represents normal urinary acidification; a failure to do so is diagnostic of dRTA.

Renal biopsy tissue was uniformly fixed in paraffin and sections were stained with hematoxylin and eosin. Immunophenotyping was performed by additional immunostaining with polyclonal antibodies to CD3, CD4, CD8, CD20, CD1a and CD138.

Where slides were available to review, we scored each biopsy to assess the nature of the inflammatory cell infiltrate and the degree of interstitial scarring. The infiltrate was scored as being either patchy or diffuse, the approximate amount of interstitium involved (<25 %, 25–50 %, 50–75 %, >75 %); the intensity of the infiltrate was scored as 1+ (light), 2+ (moderate) or 3+ (heavy); the predominant cell type in the infiltrate was recorded as was the amount of scarring, again recorded as 1–3 +.

Patients were treated with an antiproliferative agent, mycophenolate mofetil (MMF) or, in one case, azathioprine and where possible, a short reducing course of corticosteroid. The dose of MMF was titrated according to symptoms and recovery of renal function, and, in hypergammaglobulinemic patients, with the aim of bringing the IgG level within the normal range.

Statistical analysis was performed using GraphPad Prism 5.03, statistical significance was determined using the student’s t-test for Gaussian and Wilcoxon matched pair signed rank test for non-Gaussian distributed data.

The series encompasses twelve patients, 11 of these are female, with a median age of 52.5 (interquartile range (IQR) 13.5 years). At presentation the patients tended toward a metabolic acidosis (median serum bicarbonate 21 mmol/L, IQR 5.75), and had moderate renal impairment (median serum creatinine 133 μmol/L, IQR 53.2), but little or no hypokalaemia (median serum potassium 4.2 mmol/L, IQR 1.25).

9 patients had urinary acidification testing with furosemide and fludrocortisone or ammonium chloride. All but two of nine patients had abnormal urinary acidification, indicating dRTA (median baseline urine pH 6.24, IQR 1.3; median nadir urine pH 5.77, IQR 0.91); of the remaining two, one had an equivocal result (Fig. 1). A further 2 patients had a systemic acidosis and an inappropriately alkaline pH at baseline. Overall, 75 % had evidence of abnormal urinary acidification.

Proximal tubular dysfunction was assessed in 8 patients by measuring urinary retinol binding protein (RBP). The urinary RBP/Creatinine ratio was abnormally high in 5 patients (median 102 μg/mmol, IQR 1867, normal range 4–32 μg/mmol), and 2 patients had extremely high ratios (as high as 5885 μg/mmol).

All patients underwent a renal biopsy. Histology revealed a tubulointerstitial nephritis in all patients. Of the 14 biopsies performed in this series, 11 were available for us to review the tissue in addition to having the formal report.

The inflammatory cell infiltrate was diffuse in two of the biopsies and patchy in the remainder, and the intensity was rated as 1+ (light) in 5, 2+ (moderate) in 5 and 3+ (heavy) in 1. The proportion of interstitium affected by the infiltrate was <25 % in 3, 25–50 % in 2, 50–75 % in 2 and >75 % in 4. In all cases the predominant cell type was mononuclear lymphocytes. Scarring was judged to be 1+ (light) in 6, 2+ (moderate) in 4, and 3+ (heavy) in 1.

Thus, the infiltrate tended to be patchy, light to moderate in intensity and to affect the majority of the interstitium. Interstitial scarring was judged to be light to moderate (Fig. 2 for representative histology).

Three patients also had coexisting glomerular disease; one was coincidental thin basement membrane disease found on electron microscopy, the remaining two had mild glomerular mesangial matrix expansion.

Immunophenotyping revealed that the inflammatory interstitial infiltrate was a T-cell infiltrate, and was predominantly CD4+, with lesser amounts of CD8+ cells and almost no CD20+ B-cells (see Fig. 3 for representative pictures).

Of the 12 patients, 11 were treated with MMF (median dose 1000 mg/day, 1000–1500 mg), 2 were treated with azathioprine (median dose 62.5 mg/day) and 1 had no immunosuppressive treatment at all. 9 patients were treated with Prednisolone (median starting dose 10 mg/day, 5–20 mg) weaned over 3–6 months. 4 patients remain on low dose Prednisolone (median dose 7.5 mg/day, 5–10 mg) in conjunction with MMF. The median duration of immunosuppressive treatment was 24 months (IQR 24).

Both serum creatinine and eGFR improved following treatment; the median serum creatinine decreased significantly following immunosuppression treatment (153 μmol/L, IQR 25, to 124 μmol/L, IQR 39, p = 0.02), as did the eGFR (32mls/min/1.73 m², IQR 12 to 42mls/min/1.73 m², IQR 21, p = 0.04). The GFR was directly measured in 6 patients, which also improved following treatment (37mls/min/1.73 m², IQR 10.5, to 46mls/min/1.73 m², IQR 6.5, p = 0.04). Serum IgG levels decreased significantly following treatment (20.1 g/L, IQR 7.1 to 17.3 g/L, IQR 8.8, p = 0.01) (Fig. 4).

Primary Sjögren syndrome has been estimated to affect 0.6 % of the US population (1.3 million US adults) [10] and has a worldwide distribution with consistent clinical characteristics [11].

Data on the prevalence of renal disease in pSS comes from two sources; retrospective studies looking for overt renal dysfunction, mainly defects in excretory renal function and proteinuria [5, 12, 13], and smaller prospective series in which tubular defects were explicitly screened for [3, 14–16]. While the retrospective studies found a rate of renal involvement in pSS of 4.3–6.5 % [5, 12, 13], the prospective studies observed a much higher rate of 28–42 % [3, 15].

The rarer glomerular lesions, typically the characteristic membranoproliferative glomerulonephritis (MPGN), occur due to immune complex deposition following B-cell expansion. These immune complexes are often cryoglobulins [5], which bind to vascular endothelium and may cause a small vessel vasculitis. This clinical picture occurs with clonal B-cell expansion, is usually associated with lymphoma, and tends to occur late in the course of the disease [4, 5].

The more typical tubulointerstitial nephritis can manifest in a number of ways, including with a mild to moderate reduction in the glomerular filtration rate, although progression to end stage renal failure with pSS TIN is reported [5]. The tubular dysfunction can manifest as proximal tubular dysfunction with low molecular weight proteinuria [3] or a full-blown renal Fanconi syndrome [17] (with bicarbonaturia, uricosuria, phosphaturia, glycosuria and low molecular weight proteinuria). Distal tubular disease may cause the characteristic distal renal tubular acidosis [18] (with hypokalaemia, nephrocalcinosis/kidney stones, osteomalacia and a variable metabolic acidosis) or nephrogenic diabetes insipidus [19] (NDI).

These tubular lesions often require special investigations to diagnose; assays for low molecular weight proteinuria (e.g. retinol binding protein), urinary acidification tests to diagnose dRTA and water deprivation tests to diagnose NDI may not be available outside of specialist centres.

We would advocate renal biopsy in all patients with pSS and tubular defects to confirm the diagnosis of TIN, and to distinguish from other potential causes of localised tubular defects (eg. the presence of light chain).

The TIN of pSS is characterized by an invasion of mononuclear lymphocytes. We found that these cells are predominantly CD4+ T-cells with lesser populations of CD8+ T-cells; this is in keeping with data from mouse models of renal pSS [20] as well as data in human renal tissue [21], although one study suggested that CD8+ T-cells are more dominant [22].

This pattern of epithelial inflammation is strikingly similar to that seen in labial salivary glands (LSGs), which are much more intensively studied. There is evidence that the infiltrating CD4+ T-cells may be pathogenic Th17 cells; IL-17 levels are elevated in the salivary glands and serum of pSS patients and these levels correspond to the severity of the histological lesion [23]. Furthermore, knocking out IL-17 prevents disease in a mouse model of Sjögren syndrome [24]. In more extensive histological disease in LSGs, B-cells become more pronounced in the infiltrate, and it has been suggested that treatment could be tailored to the type of infiltrating cell [25].

However, there are no proven effective systemic immunosuppressive treatments for pSS; the few randomized trials have been inconclusive or contradictory.

Typical treatment for patients with joint or skin disease with pSS, is often based on hydroxychloroquine or methotrexate. Resistant or extraglandular disease has been treated variously with corticosteroids, antiproliferative agents, calcineurin inhibitors, cyclophosphamide or B-cell depletion therapy [26].

The patients in this series were all treated with MMF; the rationale being that MMF would affect both T-cell and B-cell populations and should therefore be effective across the spectrum of pSS TIN. Indeed, MMF has been used successfully with acute TIN in other settings [27]. The optimal dose of MMF to use is unknown and the dose of MMF used in our series was relatively low. We adopted a pragmatic approach and escalated the MMF in the presence of persistent symptoms, or failure of recovery of renal function. We aimed where possible to normalise the serum IgG, as there is a well-recognised association between hypergammaglobulinaemic states and dRTA (originally reported by Morris and Fudenberg [28]) and TIN, even in the context of idiopathic hypergammaglobulinaemia [29]. We also used a short weaning course of corticosteroids, if there was no contraindication to do so, as suggested by two previous series [17, 30]. This regime was well tolerated, with only one patient unable to continue with MMF due to reversible leucopenia.

We have demonstrated a significant improvement in the MDRD eGFR in the patients treated in this way; this is the first series to do so in pSS TIN alone. That the measured ⁵¹Cr-GFR in 6 of these patients significantly improved suggests that this effect is real. Two other series have shown variable improvement in excretory renal function using corticosteroids and various immunosuppressants in a cohort that included both TIN and MPGN patients [17, 30]. We suspect that specific tubular defects occur relatively early within the disease process and during the stage of active inflammation. We aim to abrogate this inflammation prior to the development of significant fibrosis, and progressive renal disease.

Patients with pSS TIN present with significant renal impairment and frequent tubular defects. The lesion is characterized by a mononuclear lymphocytic infiltrate that is predominantly composed of CD4+ T-cells and subsequent scarring. Treatment with MMF with or without a course of oral corticosteroids improved excretory renal function. Further studies into the pathogenesis and treatment of pSS TIN are warranted, given the frequency of this disease. Furthermore such studies may bring valuable insights into other inflammatory processes in the renal interstitium, such as acute allergic TIN and renal allograft rejection.