Urinary macrophages as activity markers of renal injury

Clinica Chimica Acta 297 (2000) 123–133 www.elsevier.com / locate / clinchim

Review

Urinary macrophages as activity markers of renal injury Osamu Hotta*, Naoko Yusa, Hiroshi Kitamura, Yoshio Taguma Department of Nephrology, Sendai Shakaihoken Hospital, Tsutsumimachi 3 -16 -1, Aoba-ku, Sendai 981 -8501, Japan Received 20 September 1999; received in revised form 15 December 1999; accepted 10 February 2000

Abstract The presence of macrophages (Mf) in the urine of patients with glomerulonephritis (GN) reflects the pathological events in the kidney, and we have discovered the following correlations between the Mf phenotype and the pattern of renal injury. (1) Urinary macrophage (Mf) counts increase in patients with proliferative GN, especially in the presence of active glomerular lesions (glomerular tuft necrosis, crescent, and endocapillary proliferation). In patients with hematuria, a combination of urinary Mf and T-lymphocyte counts can be used to differentiate proliferative GN from non-proliferative renal disease (hereditary nephropathy and idiopathic renal hematuria). (2) The urinary Mf of patients with active proliferative GN express FcgRIII (CD16) regardless of the type of GN. (3) There are two types of urinary CD68 1 cells, CD68 1 25F9 2 cells (infiltrating Mf) and CD68 1 25F9 1 cells (mature Mf). The CD68 1 25F9 2 cell counts in the urine correlate well with the activity of proliferative GN, and the CD68 1 25F9 1 cell counts in the urine correlate with the magnitude of non-selective proteinuria and with the subsequent decline of renal function. The CD68 1 25F9 1 cell count increases in the urine of patients with focal segmental glomerular sclerosis, but their numbers are negligible in minimal change nephrotic syndrome. These findings indicate that the analysis of the urinary Mf phenotype is a useful strategy for evaluating renal injury as a ‘risk-free renal biopsy’.  2000 Elsevier Science B.V. All rights reserved. Keywords: Urinary macrophage; Disease activity; Protein-loading tubulopathy; Lipid-laden macrophages

*Corresponding author. Tel: 1 81-22-275-3111 (ext. 358); fax: 1 81-22-275-6033. E-mail address: [email protected] (O. Hotta) 0009-8981 / 00 / $ – see front matter  2000 Elsevier Science B.V. All rights reserved. PII: S0009-8981( 00 )00239-4

124

O. Hotta et al. / Clinica Chimica Acta 297 (2000) 123 – 133

1. Introduction It is important to be able to predict progression, and estimating disease activity is important in managing renal parenchymal diseases. Renal biopsy remains the only tool for predicting the progression or evaluating the pathology underlying renal diseases, but this procedure carries some risk, making it difficult to perform frequent biopsies to monitor the status of the disease. Moreover, if the biopsy specimen is small, focal lesions such as crescent and segmental sclerosis may be overlooked [1]. We have found that the number of urinary macrophages (Mf) is increased in patients with proliferative glomerular diseases in the active stage, regardless of the type of glomerular disease [2–4], and that the phenotype of urinary Mf in glomerular disease is considerably different from that of circulating monocytes [5,6]. In this article we review the potential clinical use of an analysis of urinary Mf for monitoring the pathological status of glomerular diseases.

1.1. Flow cytometric analysis of urinary Mf Macrophages are readily identified in the glomerulus in most forms of proliferative glomerulonephritis, particularly rapidly progressive glomerulonephritis. Flow cytometry (FCM) is a powerful tool that enables precise examination of the characteristics of cells, including mononuclear cells in the urine of patients with proliferative glomerulonephritis. Unlike examination by renal biopsy, this method is non-invasive and thus facilitates serial examination [3]. Urinary Mf counts (CD14 1 cells) calculated by FCM were significantly increased in patients with proliferative glomerulonephritis who had active glomerular lesions, as demonstrated by the presence of cellular crescents, necrosis of glomerular tufts, and endocapillary proliferation, regardless of type of GN [3]. The Mf counts exceeded the T-cell counts in the urine of the patients with proliferative glomerulonephritis, whereas the Mf counts were lower than the T-cell counts in the urine of patients with hematuria-positive non-proliferative renal diseases, such as hereditary nephropathy and idiopathic renal hematuria [4]. Both the Mf count and the Mf (CD14 1 cell) / T cell (CD3 1 cell) ratio were found to be increased in proliferative glomerulonephritis, suggesting that Mf appears in the urine as a result of involvement in acute inflammation rather than via passive inflow from the bloodstream, as in the case of erythrocytes. In addition to the increased Mf excretion in the urine, the urinary Mf of active proliferative glomerulonephritis (e.g. acute phase of crescentic glomerulonephritis and post-infectious acute glomerulonephritis) expressed CD16 antigen (FcgRIII, Fc gamma receptor III) (Fig. 1), whereas the urinary Mf of nonproliferative glomerular disease and inactive proliferative glomerulonephritis did

O. Hotta et al. / Clinica Chimica Acta 297 (2000) 123 – 133

125

Fig. 1. Two-color flow cytometry analysis on urinary mononuclear cells of a patient with crescentic glomerulonephritis.

not [6]. The urinary Mf of patients with active IgA nephropathy expressed CD16 antigen, but expression was lost when patients were successfully treated with steroids (Fig. 2). CD16 1 antigen is the low-affinity receptor (FcgRIII) for IgG, which is considered to play an important role in inflammatory and anaphylactic responses [8,9]. CD16 has also been found on a small subset of circulating monocytes that make up less than 10% of the total monocyte population, or approximately 1% of all white blood cells [10,11]. A marked increase in the number of CD16 1 monocytes was observed following administration of recombinant human Mf colony-stimulating factor [12,13]. Increased CD16 1 Mf levels in the urine were observed in the majority of patients with proliferative glomerulonephritis in the active stage, but in none of the patients with non-proliferative glomerular or urological disease. Thus,

126 O. Hotta et al. / Clinica Chimica Acta 297 (2000) 123 – 133

Fig. 2. Serial changes of CD16 1 macrophages in the urine of a patient with crescentic IgA nephropathy in response to treatment.

O. Hotta et al. / Clinica Chimica Acta 297 (2000) 123 – 133

127

exposure of Mf to urine itself probably does not induce CD16 expression, because if it did, non-proliferative disease should also be associated with an increase in CD16 expression. Presumably, the CD16 1 Mf detected in the urine had converted from CD16 2 Mf to CD16 1 Mf either in the circulation or in the glomeruli. It is particularly noteworthy that expansion of the CD16 1 Mf population in the urine was often observed in patients with proliferative GN, even in cases without detectable CD16 1 monocytes in the peripheral blood. This finding suggests that Mf activation occurs at acute inflammatory sites in the glomeruli of these patients. Both the presence of a CD16 1 Mf population and an increase in total Mf counts in the urine were well correlated with the level of severity of acute inflammation in the glomeruli. Both disappearance of the CD16 1 Mf with a decrease in total Mf count in the urine and amelioration of the urinary abnormalities were brought about by steroid therapy in active proliferative glomerulonephritis patients. These findings strongly suggest that excretion of CD16 1 Mf reflects the involvement of CD16 1 Mf in the development of active proliferative glomerular lesions, such as cellular crescents, tuft necrosis and / or endocapillary proliferation, regardless of the type of glomerulonephritis. In other words, CD16 1 Mf-related tissue injury is a common event that occurs in the active stage of glomerulonephritis in many types of proliferative glomerulonephritis, regardless of the pattern of immunoglobulin deposition in the glomeruli. From a clinical viewpoint, our study suggests that detection of a CD16 1 Mf

Fig. 3. Differential diagnosis of hematuria-positive patients based on the phenotypic analysis of urinary mononuclear cells.

128

O. Hotta et al. / Clinica Chimica Acta 297 (2000) 123 – 133

population in the urine in combination with the total urinary Mf and T cell count obtained by flow cytometry is a useful non-invasive strategy for detecting active proliferative glomerulonephritis (Fig. 3).

2. Immunocytostaining of urinary Mf

2.1. Characteristics of urinary Mf ( CD68 1 cells) Immunocytostaining combined with lipid (ORO, oil red O) staining allowed the urinary CD68 (marker of monocyte / Mf)-positive cells to be divided into two subsets, ORO 1 CD68 1 cells and ORO 2 CD68 1 cells. The ORO 2 CD68 1 cells were homogeneous in size, and their size was the same as that of circulating monocytes. The ORO 1 CD68 1 cells, in contrast, were heterogeneously large. Dual staining with ORO and various monoclonal antibodies showed that most ORO 1 cells were positive for CD68, 25F9 and vimentin (Fig. 4), but that the vast majority of ORO 1 cells were negative for cytokeratin (Table 1). The phenotypical characteristics of urinary ORO 1 CD68 1 cells (lipid-laden

Fig. 4. Immunocytostaining with anti-CD68 antibody in combination with oil red O. Fatty bodies identified by oil red O are positive for CD68 antigen.

O. Hotta et al. / Clinica Chimica Acta 297 (2000) 123 – 133

129

Table 1 Phenotype of urinary macrophages

CD14 CD16 (FcgRIII) CD68 Class (HLA-DR) 25F9 Lipids Cytokeratin

Activated Mf

Lipid-laden Mf

1 1 1 1 2 2 2

2 2 1 1 1 1 2

macrophages) were the same as ‘giant Mf’ observed in biopsy specimens, i.e. they were positive for CD68, 25F9 (mature Mf), HLA-DR, CD71 (transferrin receptor) and negative for KL1 (pan cytokeratin) [7]. Giant Mf in the kidney biopsy specimens were mainly distributed within tubular lumens and Bowman’s space. Thus, the detection of ORO 1 CD68 1 cells in the urine indicates the presence of intertubular giant Mf in the kidneys. The presence of intertubular giant Mf were restricted to progressive glomerular diseases, such as rapidly progressive glomerulonephritis, IgA nephropathy (especially advanced form), collapsing glomerulonephropathy / focal segmental glomerulosclerosis, diabetic nephropathy, membranoproliferative glomerulonephritis, membranous nephropathy, and Alport’s syndrome [7,14]. In contrast, giant Mf were not observed in non-progressive forms of glomerular disease, including minimal change nephritic syndrome, the early stages of IgA nephropathy, and thin basement membrane disease [7]. Lipids are found in three forms in the urine in renal disease: free fat droplets, entrapped in casts, and oval fatty bodies. Oval fatty bodies have been thought to represent tubular epithelial cells that have undergone lipid degeneration [15]. However, our results showed that the fatty bodies in the urine expressed Mf-associated antigens and for the most part lacked cytokeratin-associated antigen. Although the mechanism by which ORO 1 CD68 1 cells appear in the urine is uncertain at present, there are several possible mechanisms for the appearance fatty bodies (ORO 1 CD68 1 cells) in the urine. Because most of the ORO 1 cells are positive for Mf antigens (CD68 and 25F9), but negative for an epithelial cell marker (cytokeratin), the most simple explanation is that urinary fatty bodies (ORO 1 CD68 1 cells) are of Mf origin and not of tubular epithelial cell origin. However, Bariety et al. hypothesized that podocytes detached from the glomerular basement membrane or tubular cells may undergo metaplastic changes into macrophage-like cells that are the same as giant Mf [14]. Experimentally, it is now known that glomerular epithelial cells transform into

130

O. Hotta et al. / Clinica Chimica Acta 297 (2000) 123 – 133

macrophage-like cells in normal rat glomerular cultures and that the change is enhanced by colony-stimulating factors [16]. Further study is needed to determine the origin of giant Mf (or fatty bodies) both in kidneys and in urine.

2.2. Clinical features and urinary CD68 1 cells No ORO 1 CD68 1 cells were detected in the urinary sediments of the normal subjects. The number of ORO 2 CD68 1 cells was negligible in the most of the normal subjects, but a small number of ORO 2 CD68 1 cells ( , 50 cells / 5 ml urine) were present in the urinary sediment of some normal cases. The number of ORO 2 CD68 1 cells and / or ORO 1 CD68 1 cells in the urinary sediment of patients with renal parenchymal disease were increased to various degrees. A marked increase in ORO 2 CD68 1 cells was observed in the acute phase of proliferative glomerulonephritis, the same as the increase in CD14 1 cells identified by flow cytometry. Significant urinary ORO 1 CD68 1 cell excretion was observed in the patients with non-selective proteinuria, but the number of ORO 1 CD68 1 cells was negligible in the urine of patients without proteinuria, even if they had severe hematuria. The ORO 1 CD68 1 cells were also negligible or only transiently detectable in the urine of the patients with minimal change nephrotic syndrome. In contrast, an increase in urinary ORO 1 CD68 1 cells was observed to various degrees in patients with non-selective proteinuria, regardless of the type of glomerular disease, including rapidly progressive glomerulonephritis, focal segmental sclerosis, IgA nephropathy, membranoproliferative glomerulonephritis, membranous nephropathy, and Alport’s syndrome. Thus, detection of urinary ORO 1 CD68 1 cells is a good indicator of differential diagnosis such as focal segmental glomerular sclerosis vs. minimal change nephritic syndrome, and Alport’s syndrome vs. thin basement membrane disease. The number of urinary ORO 1 CD68 1 cells was significantly correlated with the magnitude of proteinuria (Fig. 5). Furthermore, prolonged, marked urinary ORO 1 CD68 1 cell excretion was usually associated with subsequent deterioration of renal function, and the number of ORO 1 CD68 1 cells in the urine was significantly correlated with the degree of the subsequent decline of renal function (Fig. 6). Although the origin of urinary ORO 1 CD68 1 cells has not yet been fully elucidated, they may be derived from mature infiltrating Mf or metaplastic epithelial cells. Our results indicate that urinary ORO 1 CD68 1 cell counts may be used as an indicator of non-selective proteinuria-induced tubular injury. In summary: (1) urinary Mf counts are increased in patients with proliferative GN, especially in the presence of active glomerular lesions (glomerular tuft necrosis, crescent, and endocapillary proliferation). In patients with hematuria, a

O. Hotta et al. / Clinica Chimica Acta 297 (2000) 123 – 133

131

Fig. 5. Relationships between the urinary lipid-laden CD68 1 cell counts and magnitude of proteinuria.

Fig. 6. Relationships between the urinary lipid-laden CD68 1 cell counts and decline in renal function identified by reciprocal serum creatinine level.

132

O. Hotta et al. / Clinica Chimica Acta 297 (2000) 123 – 133

combination of urinary Mf and T-lymphocyte counts can be used to differentiate proliferative GN from non-proliferative renal disease (hereditary nephropathy and idiopathic renal hematuria). (2) The urinary Mf of patients with active proliferative GN express FcgRIII (CD16) regardless of the type of GN. (3) There are two types of urinary CD68 1 cells, ORO 2 CD68 1 25F9 2 cells (infiltrating Mf) and ORO 1 CD68 1 25F9 1 cells (mature Mf, the same as fatty bodies). The ORO 2 CD68 1 25F9 2 cell counts in the urine correlate well with the activity of proliferative GN, and the ORO 1 CD68 1 25F9 1 cell counts in the urine correlate with the magnitude of non-selective proteinuria and with the subsequent decline of renal function. These findings indicate that the analysis of the urinary Mf phenotype is a useful strategy for evaluating renal injury as a ‘risk-free renal biopsy’.

References [1] Hotta O, Taguma Y, Sudo K, Kurosawa K. Limitation of kidney biopsy in detecting crescentic lesions in IgA nephropathy. Nephron 1993;65:472–3. [2] Hotta O, Taguma Y, Ooyama M, Yusa N, Nagura H. Analysis of CD14 1 cells and CD56 1 cells in urine using flow cytometry: a useful tool for monitoring disease activity of IgA nephropathy. Clin Nephrol 1993;39:289–94. [3] Hotta O, Taguma Y, Yusa N, Ooyama M. Analysis of mononuclear cells in urine using flow cytometry in glomerular diseases. Kidney Int 1994;47(Suppl.):117–21. [4] Hotta O, Yusa N, Ooyama M, Taguma Y. Urinary macrophage counts and ratio to T lymphocytes: possible use in differential diagnosis and management of glomerular disease. J Clin Lab Anal 1996;10:205–8. [5] Hotta O, Kitamura H, Taguma Y. Detection of mature macrophages in urinary sediments: clinical significance in predicting progressive renal disease. Renal Failure 1998;20:413–8. [6] Hotta O, Yusa N, Ooyama M, Unnno K, Furuta T, Taguma Y. Detection of urinary macrophages expressing the CD16 (FcgRIII) molecule: a novel marker of acute inflammatory glomerular injury. Kidney Int 1999;55:1927–34. [7] Oda T, Hotta O, Taguma Y et al. Clinicopathological significance of intratubular giant macrophages in progressive glomerulonephritis. Kidney Int 1998;53:1190–200. [8] Weinshank RL, Luster AD, Ravetch JV. Function and regulation of a murine macrophagespecific IgG Fc receptor, FcgR-a. J Exp Med 1988;163:1909–25. [9] Hazenbos WLW, Gessner JE, Hofhuis FMA et al. Impaired IgG-dependent anaphylasis and Arthus reaction in FcgRIII (CD16) deficient mice. Immunity 1996;5:181–8. [10] Passlick B, Flieger D, Ziegler-Heitbrock HWL. Identification and characterization of novel monocyte subpopulation in human peripheral blood. Blood 1989;74:2527–34. ¨ [11] Ziegler-Heitbrock HWL, Strobel M, Kieper D et al. Differential expression of cytokines in human blood monocyte subpopulations. Blood 1992;79:503–11. [12] Saleh MN, Goldman SJ, LoBuglio AF et al. CD16 1 monocytes in patients with cancer: spontaneous elevation and pharmacologic induction by recombinant human macrophage colony-stimulating factor. Blood 1995;85:2910–7. [13] Schmid I, Baldwin GC, Jacobs EL et al. Alterations in phenotype and cell-surface antigen expression levels of human monocytes: differential response to in vivo administration of rhM-CSF or rhGM-CSF. Cytometry 1995;22:103–10.

O. Hotta et al. / Clinica Chimica Acta 297 (2000) 123 – 133

133

[14] Bariety J, Nochy D, Mandnet C, Jacquot C, Glotz D, Meyrier A. Podocytes undergo phenotypic changes and express macrophagic associated markers in idiopathic collapsing glomerulopathy. Kidney Int 1998;53:918–25. [15] Hudson JB, Dennis A, Gerhardt RE. Urinary lipid and the Maltese cross. New Engl J Med 1978;299:586. [16] Orikasa M, Iwanaga T, Takahashi-Iwanaga H, Kozima K, Shimizu F. Macrophagic cells outgrowth from normal rat glomerular culture: possible metaplastic change from podocytes. Lab Invest 1996;75:719–33.