PATHOPHYSIOLOGY of the RENAL BIOPSY www.jasn.org

Immune Complex Tubulointerstitial Nephritis Due to Autoantibodies to the Proximal Tubule Brush Border

† ‡ Ivy A. Rosales,* A. Bernard Collins,* Paula Alves S. do Carmo, Nina Tolkoff-Rubin, R. Neal Smith,* and Robert B. Colvin*

Department of *Pathology and ‡Division of Nephrology, Renal Transplantation Program, Massachusetts General Hospital, Boston, Massachusetts; and †Centro Especializado em Anatomia Patologica, Belo Horizonte, Minas Gerais, Brazil

ABSTRACT Immune complex tubulointerstitial nephritis due to antibodies to brush border associated with focal tubulitis. Interstitial antigens of the proximal tubule has been demonstrated experimentally and rarely in fibrosis and tubular atrophy were diffuse humans. Our patient developed ESRD and early recurrence after transplantation. and present in 70% of the cortex. No gi- IgG and C3 deposits were conspicuous in the tubular basement membrane of antcellswereseen.Anarteryshowed proximal tubules, corresponding to deposits observed by electron microscopy. mild intimal fibrosis. Few IgG4-positive Rare subepithelial deposits were found in the glomeruli. The patient had no plasma cells were detected in the intersti- evidence of SLE and had normal complement levels. Serum samples from the tium (mean of 2 per high-power field). patient reacted with the brush border of normal human kidney, in contrast with the Immunofluorescence (IF) showed negative results with 20 control serum samples. Preliminary characterization of widespread granular deposits along the the brush border target antigen excluded megalin, CD10, and maltase. We postulate proximal tubular basement membrane that antibodies to brush border antigens cause direct epithelial injury, accumulate in (TBM), which stained for IgG (3+), C3 the tubular basement membrane, and elicit an interstitial inflammatory response. (3+), and equally for k and l (Figure 1). The TBM deposits stained with IgG1 J Am Soc Nephrol 27: 380–384, 2016. doi: 10.1681/ASN.2015030334 (widespread 3+), IgG2 (focal 2+), and IgG4 (widespread 3+). IgG3 was nega- tive. The glomerular deposits were sparse, and they showed very segmental CLINICAL HISTORY blood cell counts and platelets were nor- staining for IgG and did not stain for the mal. Serologies were negative, including phospholipase-A2 receptor. A 73-year-old man presented with nausea, ANCA, anti–glomerular basement Ultrastructural analysis showed wide- vomiting, and progressive renal failure. membrane (GBM), anti-nuclear anti- spreadamorphouselectron-densedeposits His serum creatinine had increased from body, anti–double-stranded DNA, and within the TBM, sometimes abutting the 1.3 mg/dl to 15.2 mg/dl in the 10 months hepatitis B and C. C3 and C4 levels basal plasma membrane of the tubular before admission. He had a history of were normal. An ultrasound showed cell (Figure 1). Scattered penetrating coronary artery disease, diabetes, and no evidence of a mass, obstruction, or subepithelial deposits were present in a hypertension, and he was taking multi- stenosis. small minority of glomerular capillaries ple medications including omeprazole, without spike formation (Figure 1). The lisinopril, metformin, simvastatin, meto- GBM was diffusely but mildly thickened prolol, aspirin, and insulin. On admission, KIDNEY BIOPSY his BP was 148/78 mmHg. His other vital signs were normal and his physical exam- Ten histologically normal glomeruli were ination was unremarkable. He had no present with no endocapillary hypercellu- Published online ahead of print. Publication date prior history of autoimmune disease and larity or GBM abnormality. There was available at www.jasn.org. no family history of autoimmune or renal prominent diffuse tubular injury mani- diseases. fested by loss of brush borders, blebs, Correspondence: Dr. Ivy A. Rosales, Department of Pathology, Massachusetts General Hospital, 55 Urinalysis showed protein (2+), vacuoles, and regenerative changes. An Fruit Street, Thier 831A, Boston, MA 02114. Email: blood (1+), 3–5 white blood cells per interstitial infiltrate of lymphocytes, [email protected] fi high-power eld, and no casts. His he- plasma cells, and a few eosinophils af- Copyright © 2016 by the American Society of moglobin was 12.6 g/L, although white fected 50%–60% of the cortex (Figure 1), Nephrology

380 ISSN : 1046-6673/2702-380 J Am Soc Nephrol 27: 380–384, 2016 www.jasn.org PATHOPHYSIOLOGY of the RENAL BIOPSY

Figure 1. Native kidney.(A) Mononuclear inflammatoryinterstitial infiltrate with few admixedneutrophilsand eosinophils, tubular atrophy, and normal-looking glomerulus (hematoxylin and eosin staining). (B) Granular IgG staining of TBMs. (C) Amorphous electron-dense deposits along the TBM. (D) Penetrating subepithelial electron-dense deposits along the GBM. Original magnification, 3400 in A and B; 38900 in C and D.

(harmonic mean thickness of 602 nm), Lupus serologies remained negative and were detected in the TBM abutting the indicating early diabetic glomerulopathy. complement levels were normal. An allo- proximal tubule basal membrane, similar These findings were interpreted as graft biopsy showed a sparse inflamma- to the native kidney biopsy. These findings immune complex tubulointerstitial ne- tory infiltrate of about 5% of the cortex, were interpreted as recurrent immune phritis (TIN) with minor involvement of composed of lymphocytes, occasional complex TIN. the glomerulus resembling early mem- neutrophils, and rare eosinophils. Loss The patient’s pretransplant and post- branous GN. of brush borders, dilation, and protein transplant sera were tested for anti–brush casts were focally present, but there was border autoantibodies (ABBAs) by indi- no tubulitis. Glomeruli and vessels were rect IF on cryostat sections of normal kid- CLINICAL FOLLOW-UP unremarkable. IF was similar to the native neys at a 1:5 dilution. Strong and selective biopsy, with diffuse granular TBM depos- binding of IgG to the brush border of The patient was treated with steroids but its that stained for IgG (4+), C3 (4+), and proximal tubules was evident without re- remained hemodialysis dependent. He C4d (4+). The glomeruli were negative. activity to the TBM (Figure 2). Control received a deceased donor allograft kidney C4dwasnegativeinperitubularcapillaries. normal human sera (n=20) did not stain 4 years later. The immediate post-transplant The polyoma stain was negative (SV40 brush borders under the same conditions. course was unremarkable and he was large Tantigen). The TBMdeposits stained To seek the antigenic target, double staining maintained on tacrolimus, mycophenolate forIgG1(widespread3+)andwerenega- with the patient’s serum was repeated com- mofetil, and prednisone. tive for IgG2, IgG3, and IgG4. The signif- bined with antibodies to megalin (Figure 3), Seven weeks later, the patient again icance of the difference in IgG subclasses CD10, or maltase. No colocalization was developed nausea and vomiting and his in the native and allograft biopsies is not present (data not shown). serum creatinine level increased (2 mg/dl). clear. Amorphous electron-dense deposits Plasma exchange was initiated, along with intravenous Ig and rituximab, without dramatic benefit. At 45 weeks post-transplant, the patient’sserum creatinine was 3 mg/dl. A repeat allo- graft biopsy showed progression of the TIN, now involving 30% of the cortex, loss of tubular brush borders, and flat- tened epithelial cells and persistent TBM deposits by IF and electron mi- croscopy (Figure 4). C4d was negative in peritubular capillaries and SV40 was negative. Rare subepithelial glomerular deposits were identified. The patient received methylprednisolone pulses, Figure 2. Indirect IF of the patient’s serum on normal human kidney. (A) IgG staining of rituximab, and later belatacept. Serum proximal tubule brush borders. (B) Proximal tubule brush border IgG staining and negative creatinine levels slowly increased to glomerular staining. (C) Control normal serum on normal human kidney. 3.1 mg/dl 2 years post-transplant.

J Am Soc Nephrol 27: 380–384, 2016 Anti–Brush Border Antibodies and Nephritis 381 PATHOPHYSIOLOGY of the RENAL BIOPSY www.jasn.org

failure.3 The antigenic target remains uncertain, but ABBAs have not been reported. Idiopathic hypocomplementemic TIN is manifested by granular TBM de- posits of IgG and C3 in association with hypocomplementemia. This preferential TBM deposition has been reported in eight cases, all in the setting of low com- plement levels and not associated with SLE or Sjögren’s syndrome. ABBAs have not been reported in this setting.4 Drug-induced TIN with TBM im- Figure 3. Patient IgG-megalin on normalhuman kidney, colocalization by IF. (A) Patient IgG mune complex deposition has been staining on proximal tubule brush borders. (B) Megalin staining on proximal tubule brush rarely documented in association with borders. (C) No evident IgG-megalin colocalization. nonsteroidal anti-inflammatory drugs.5 Polyclonal IgG and C3 TBM deposits are Antibody titers were determined by should be considered in the differential present without glomerular deposits. indirect IF using sera from pretransplant diagnosis (Table 1). Two cases showed giant cell tubulitis and post-transplant immunosuppressive Lupus nephritis is the most common with TBM deposits.6 In both settings, treatment time points. The titer before cause of immune complex deposition in the deposits were postulated to arise kidney transplantation was .1:960. Be- the TBM, occurring in about 60% of from antidrug antibodies or elicited fore plasma exchange, the titer was noted lupus GN. Isolated TIN in SLE has also autoantibodies. at =1:240 and was at =1:120 ten weeks been reported.1,2 Polyomavirus nephropathy can cause after plasma exchange and immunosup- IgG4-related systemic disease (IgG4- focal granular TBM deposits of IgG, C3, pressive therapy (intravenous Ig, rituximab, RSD) typically presents with multiorgan and C4d in addition to TIN and viral tacrolimus, and steroids). There was an involvement, including autoimmune pan- cytopathic changes, most often seen in apparent decrease in antibody titer after creatitis, inflammatory bowel disease, scle- allograft kidneys. The nature of the transplantation, along with some decrease rosing cholangitis, and/or sialadenitis, as well antigen is uncertain. These have been in titer after plasma exchange and immu- as TIN. Hypergammaglobulinemia with el- reported to react with one antiserum but nosuppressive therapy. evated IgG4 levels is usually present and not with antibodies to major capsid 50% of patients have hypocomplementemia. protein, VP1, minor capsid proteins, Renal biopsy shows IgG4+ plasma cell– VP2 and VP3, or large T antigen.7,8 DISCUSSION rich infiltrate, storiform fibrosis, and The clinical, laboratory, and biopsy granular TBM immune complex depos- findings in our patient rule out these There are several causes of immune itsforIgGandC3.IgG4-RSDissteroid more common causes of immune com- complex–mediated TIN in humans that responsive and rarely causes renal plex TIN. In this patient, there was no

Figure 4. Allograft kidney. (A) Tubular epithelial blebs, vacuolization, focal tubulitis, and mononuclear inflammatory interstitial infiltrate, 45 weeks post-transplant (hematoxylin and eosin staining). (B) Granular IgG staining of TBMs, 45 weeks post-transplant. (C) Large amorphous electron-dense depositsalongtheTBM,7 weeks post-transplant;similar deposits are presentat 45weeks.(D)Rare amorphoussubepithelial electron-dense deposits along the GBM, 45 weeks post-transplant. Original magnification, 3400 in A and B; 314,000 in C; 318,000 in D.

382 Journal of the American Society of Nephrology J Am Soc Nephrol 27: 380–384, 2016 www.jasn.org PATHOPHYSIOLOGY of the RENAL BIOPSY

Table 1. Pathologic characteristics and outcome of immune complex tubulointerstitial nephritides Electron Glomerular Disease Light Microscopy IF ABBA Outcome Microscopy Involvement Lupus nephritis Tubulitis and interstitial Granular TBM for Electron-dense Rarely with minimal 2 Partial to complete inflammation IgGs and TBM deposits glomerular response to complement involvement high-dose corticosteroids1,2 IgG4-related Storiform fibrosis with Granular TBM for Electron-dense With or without 2 General response systemic disease IgG4+plasma IgG and C3, TBM deposits membranous to steroids; cell–rich infiltrate commonly IgG4 glomerulopathy refractory subset cases may respond to rituximab3 Idiopathic Tubulitis and interstitial Granular TBM Electron-dense 22Partial to hypocomplementemic inflammation for IgG and C3 TBM deposits complete TIN response to immunosuppression4 Drug-induced TIN Tubulitis and interstitial Granular TBM for Electron-dense 22Responsive to with TBM inflammation with IgG and C3 TBM deposits steroids5 immune complex eosinophils with or deposits without neutrophils Polyomavirus Viral cytopathic Granular TBM Electron-dense 22Responsive to nephropathy changes staining for TBM deposits decrease in C4d, IgG, immunosuppression and C3 plus cidofovir, leflunomide, or intravenous Ig8 ABBA-TIN Tubular injury and Granular TBM Electron-dense Minimal + Developed ESRD over interstitial staining for TBM deposits subepithelial years; recurred in inflammation C4d, IgG, and C3 deposits transplant evidence of lupus, hypocomplementemia, incidental finding, we tested 20 control rats react with megalin in the brush border orIgG4-RSD, and there wasnoestablished sera from samples submitted for either of proximal tubules and podocytes.13 association with drugs or virus and no ANCA or anti-GBM assays. These were Passive transfer of antibodies causes de- response to steroids. We therefore believe uniformly negative for ABBA, although position of antibodies in vivo along the that this is a distinctive form of primary 15% had anti-nuclear antibodies. GBM in an epimembranous pattern, the immune complex TIN associated with Hence, we believe the ABBAs detected proximal tubule brush border and de- ABBAs. in this patient are a distinctive finding. posit in the TBM,14 with associated dis- We were able to find only one some- An experimental model for ABBA-TIN ruption of the normal architecture of what similar case in the literature.9 A was reported by Klassen et al. in 1977. the proximal tubules.14,15 In humans, patient with myasthenia gravis devel- They immunized rabbits with nonglo- megalinisexpressedinthebrushbor- oped ARF owing to TIN, granular IgG merular components of the rabbit kidney, derbutnotinpodocytes.ABBAshave and C3 deposits in the proximal TBM, which led to tubular injury, interstitial fi- also been noted in murine chronic graft- and epimembranous deposits in the brosis, and focal lymphocytic infiltrates versus-host disease.16 GBM. ABBAwas documented in the pa- with granular IgG and C3 deposition The native kidneys in this patient failed tient’s serum and in the eluate from the along the TBM of proximal tubules.12 An- because of progressive tubulointerstitial post mortem kidney. ABBAs have also tibodies eluted from diseased kidneys inflammation and scarring, which sug- been detected rarely in membranous reacted with the proximal tubule brush gests persistent antibody formation. This GN, but TBM deposits were not noted.10 border and the TBM deposits.12 The na- hypothesis was supported by persistence In the older literature, ABBAs have ture of the antigen was not further char- of the ABBA and early recurrence in the been reported in Crohn’sdiseaseandin acterized. This study may provide insight allograft. We postulate that IgG autoanti- about 10% of sera submitted for auto- into the pathophysiology of ABBA in our body to an unknown antigen on the antibody screening; however, these were current case. However, the trigger of the proximal tubule brush border likely not kidney specific and were not associ- ABBA in our case is unknown. caused direct epithelial cell injury, accu- ated with renal disease.11 To address the Another experimental model is mulated as in situ immune complexes and possibility that the ABBA was an Heymann nephritis. Sera from these caused interstitial inflammation and

J Am Soc Nephrol 27: 380–384, 2016 Anti–Brush Border Antibodies and Nephritis 383 PATHOPHYSIOLOGY of the RENAL BIOPSY www.jasn.org

fibrosis. Rare subepithelial deposits in the 4. Kambham N, Markowitz GS, Tanji N, circulating antibodies against brush border ’ native and transplant kidney suggest that Mansukhani MM, Orazi A, D Agati VD: Idio- antigens (Fx1A) in a patient with membranous pathic hypocomplementemic interstitial nephropathy and bilateral pyeloureteral ste- the antigen is expressed to a minor degree nephritis with extensive tubulointerstitial nosis. Comparison with idiopathic membra- by the podocyte. deposits. Am J Kidney Dis 37: 388–399, 2001 nous nephropathy. Nephrol Dial Transplant 7: 5. Dixit MP, Nguyen C, Carson T, Guedes B, 293–299, 1992 Dixit NM, Bell JM, Wang Y: Non-steroidal 11. Skogh T, Heuman R, Tagesson C: Anti-brush anti-inflammatory drugs-associated acute in- border antibodies (ABBA) in Crohn’sdisease. ACKNOWLEDGMENTS terstitial nephritis with granular tubular base- J Clin Lab Immunol 9: 147–150, 1982 ment membrane deposits. Pediatr Nephrol 12. Klassen J, Milgrom FM, McCluskey RT: This work was made possible through an 23: 145–148, 2008 Studies of the antigens involved in an im- Am International Society of Nephrology Fellow- 6. Chang A, Peutz-Kootstra CJ, Kowalewska J, munologic renal tubular lesion in rabbits. Logar CM, Gitomer JJ, Davis CL, Shankland JPathol88: 135–144, 1977 ship Grant to I.A.R. SJ, Alpers CE, Smith KD: Giant cell tubulitis 13. Kerjaschki D, Farquhar MG: Immunocyto- with tubular basement membrane immune chemical localization of the Heymann nephritis deposits: A report of two cases after cardiac antigen (GP330) in glomerular epithelial cells DISCLOSURES valve replacement surgery. Clin J Am Soc of normal Lewis rats. JExpMed157: 667–686, Nephrol 1: 920–924, 2006 1983 None. 7. Hever A, Nast CC: Polyoma virus nephropa- 14. Mendrick DL, Noble B, Brentjens JR, Andres thy with simian virus 40 antigen-containing GA: Antibody-mediated injury to proximal Kidney Int REFERENCES tubular basement membrane immune com- tubules in Heymann nephritis. 18: plex deposition. Hum Pathol 39: 73–79, 2008 328–343, 1980 8. Bracamonte E, Leca N, Smith KD, Nicosia RF, 15. Noble B, Andres GA, Brentjens JR: Passively 1. Dhingra S, Qureshi R, Abdellatif A, Gaber LW, Nickeleit V, Kendrick E, Furmanczyk PS, transferred anti-brush border antibodies in- Truong LD: Tubulointerstitial nephritis in sys- Davis CL, Alpers CE, Kowalewska J: Tubular duce injury of proximal tubules in the ab- temic lupus erythematosus: Innocent bystander basement membrane immune deposits in sence of complement. Clin Exp Immunol 56: or ominous presage. Histol Histopathol 29: association with BK polyomavirus nephrop- 281–288, 1984 553–565, 2014 athy. Am J Transplant 7: 1552–1560, 2007 16. Bruijn JA, Hogendoorn PC, Corver WE, van 2. Singh AK, Ucci A, Madias NE: Predominant 9. Morrison EB, Kozlowski EJ, McPhaul JJ Jr: den Broek LJ, Hoedemaeker PJ, Fleuren GJ: tubulointerstitial lupus nephritis. Am J Kidney Primary tubulointerstitial nephritis caused by Pathogenesis of experimental lupus nephritis: Dis 27: 273–278, 1996 antibodies to proximal tubular antigens. Am A role for anti-basement membrane and anti- 3. Cornell LD: IgG4-related kidney disease. J Clin Pathol 75: 602–609, 1981 tubular brush border antibodies in murine Curr Opin Nephrol Hypertens 21: 279–288, 10. González-Cabrero J, de Nicolas R, Ortíz A, chronic graft-versus-host disease. Clin Exp 2012 Mampaso F, Hernando L, Egido J: Presence of Immunol 79: 115–122, 1990

384 Journal of the American Society of Nephrology J Am Soc Nephrol 27: 380–384, 2016