Lipiduria – with special relevance to Fabry disease

Introduction

Though organic lipids were once defined as “naturally occurring fat-like substances – soluble in organic solvents but not in water” [1] the inclusion of lipids with polar moieties, hence soluble in water, lead to the refined definition “actual or potential derivatives of free fatty acids and their metabolites” [2]. However, it is the lipids that are insoluble in water that form particles in human urine; hence we will confine this discussion to these compounds. Such lipid particles are not normally seen in the urine, and this paper will attempt to provide answers to the following questions regarding lipiduria: Does lipiduria reflect tissue histology or pathology? How is lipiduria detected by urinary microscopy? Can lipiduria cause confusion in urinary microscopy? How is this relevant to Fabry disease?

Lipids in tissue histology

To process tissues for conventional microscopy the fact that the tissue undergoes a series of “insults” is often forgotten. Fixation in either an aqueous or solvent based fixative is usually the first step, to prevent deterioration by autolysis and to increase tissue rigidity for later sectioning. These fixatives alter the natural characteristics of the tissue and may alter the nature of contained lipids and lipoproteins. The main alternative is freezing, but frozen sections ultimately result in quite poor histological definition, with a finite useful lifespan for further study. Microwave fixation, as often used in immunohistochemistry, utilises heat, which is damaging to lipids. The chemically fixed tissue is then imbedded by immersion in a fluid which sets – usually paraffin. This involves heating the tissue and the use of agents, such as graded alcohols, xylene or similar lipid solvents, to allow penetration of the paraffin wax. The section is then cut, and the paraffin removed by de-waxing with lipid solvents, a process which once again removes lipids but is helpful to allow later aqueous stains to penetrate. Finally aqueous stains, such as haematoxylin and eosin, are used to demonstrate the histology of the dewaxed, lipid removed, sections [3]. It should be no surprise that lipids in tissues are then represented by “vacuoles” or “cholesterol clefts”. The lipids are not seen – only the spaces they once occupied.

If lipids are to be sought histologically, e.g. to be sure the vacuoles seen in fatty liver cells actually contain fat, fresh frozen tissue is imbedded in an agent that requires little heat and is water soluble (such as Carbowax™, Dow Chemicals or Tissue Tek™, Fisher Scientific), and then stained with such agents as Sudan or Oil red which preferentially bind to lipids and fats [2]. The resulting tissue definition is poor.

One of the odd characteristics of cholesterol esters in fresh frozen tissue (or fluids) is that when observed under polarised light, they display distinctive double refraction resulting in the appearance of a “Maltese” or “Amalfi” cross, so named because of the resemblance to the emblem of the mediaeval Knights Hospitaller of St John. This birefringence, scientifically termed “conic focal anisotropism” is of great importance to detection of lipids in urine [4].

There are renal diseases where lipid “seen” in tissue sections (or actually not seen in conventionally processed tissues – though the vacuoles or clefts remain) has been of significant diagnostic value. In Minimal Change disease where lipid vacuoles are present in tubular cells, associated with heavy proteinuria but minimal glomerular abnormality, this appearance once lead to the diagnostic term “Lipoid nephrosis”. With cholesterol embolic disease of the kidney, diagnostic clefts are left in vessels where the cholesterol was once lodged. Fabry disease is a genetically acquired condition in which the cytoplasm of glomerular podocytes and other cells is packed with glycosphingolipid particles, leading to a foamy appearance with conventional microscopy [5]. First described simultaneously and independently by Johannes Fabry (1860–1930) and William Anderson (1842–1900) in 1898 – depending on where you are in the world it bears either Fabry alone or both as eponymous names, or the term “angiokeratosis corporis diffusum” because of the characteristic angiokeratotic skin “rash” seen in the peri-umbilical and lower torso region in many cases. It is due to an X-linked deficiency of the action of the enzyme α-galactose A, resulting in lysosomal accumulation of the glycosphingolipid ceramide trihexoside, now called globotriaosylceramide (GL3). The clinical picture varies widely with GL3 deposits particularly troublesome in skin (angiokeratosis), kidney (albuminuria, haematuria, and renal failure), nerve vessels (neuropathy, both peripheral and autonomic), heart (arrhythmia, ischaemia and cardiomyopathy) and cerebral vessels (transient ischaemic attacks, strokes and micro-infarcts). This variability in phenotype results in late diagnosis in many instances, which is particularly unfortunate now that enzyme replacement therapy, effective in delaying or preventing these complications, is now available [6, 7].

Renal biopsy is one of the methods by which the condition has been diagnosed, with lipid vacuoles in fixed section cells (foam cells), Maltese cross appearance under polarised light in frozen sections and electron microscopy demonstrating very characteristic whorled myelin bodies in the cytoplasm of cells especially glomerular podocytes, but renal biopsy is invasive, expensive and has risks for the patient.

Detection of lipid particles in urine

As no fixation/imbedding/sectioning/dewaxing is required, fresh urine specimens still retain any lipid particles, and these can easily be seen exhibiting the characteristic Maltese cross appearance in polarised light [4, 8]. Their presence has been well described in association with heavy proteinuria including Minimal Change disease but also other albuminuric conditions. Lipiduria may also be occasionally seen in polycystic kidney disease and can be seen associated with haematuria, where it has been suggested that the lipids arise from degraded red cell membranes [9]. Lipiduria can also occur in disorders, such as chyluria [10, 11], with lipid emulsion therapy associated with intra-abdominal trauma [12] as well as with prostatitis and oil contamination of urine specimens [13].

In nephrotic syndrome lipids in urine may be seen as free floating fat droplets (Figure 1), packed into renal tubular epithelial cells or macrophages (oval fat bodies) (Figure 2) and within tubular casts (fatty casts) (Figure 3). The uncommonly seen cholesterol crystals, flat plates (Figure 4), interestingly do not demonstrate birefringence hence no Maltese cross appearance. In Fabry disease, cells packed with GL3 vacuoles, “Mulberry cells”, have long been known to be characteristic [14–16], however, they are usually considered insufficiently common to use as a sensitive diagnostic tool.

Figure 1:

Fatty droplets seen in phase contrast (A) and with polarised light (B), hence Maltese cross appearances.

(From reference [8], reproduced with permission).

Figure 2:

Oval fat bodies in phase contrast (A) and polarised light (B).

(From reference [8], reproduced with permission).

Figure 3:

Fatty cast in phase contrast (A) and polarised light (B).

(From reference [8], reproduced with permission).

Figure 4:

Cholesterol crystal – phase contrast.

(From reference [8], reproduced with permission).

Confusion with lipid particles in urine

With conventional microscopy there are a variety of bodies which may be confused with lipid particles [13]. Haemoglobin – free red cells (ghosts) may appear similar to large fat droplets but they do not exhibit birefringence with polarised light. Candida, free nuclei and other cell debris may look like lipid droplets but again do not form Maltese crosses under polarised light. However, starch particles as may fall from latex gloves do exhibit birefringence and a Maltese cross appearance, but their conventional light microscopy appearance is quite different as they are irregular in size and shape and often have a central dimple.

Urine microscopy in Fabry disease

As mentioned, Fabry disease varies tremendously, and as has been recently realised, can also cause disease in the heterozygous females who have the abnormal gene on one of their X chromosomes. It can thus be difficult to detect or diagnose. Enzyme and genetic testing is expensive, slow and not available in some circumstances. Mulberry cells in the urine are helpful but uncommon. We therefore asked the question “can urine microscopy help, across a wide phenotypic spectrum?” [17].

Over decades, a large cohort of Fabry patients had been collected, initially to document the manifestations of the disease in our environment [18, 19], and also in anticipation of the development of enzyme replacement therapy, which fortunately has eventuated [6]. This gave us an opportunity to examine the urine of 29 patients, deliberately selected across a wide phenotype, all enzyme proven, most genotyped. We compared examination of their urine with 21 control subjects: 20 with other renal disease, one normal. Of the Fabry patients, nearly half (46%) were male, ages ranged from 18 to 67 years and 51% were already on enzyme replacement therapy. From a renal point of view they also varied widely. In total four females and two males had normal GFR and albumin excretion, 15 had a urinary albumin/creatinine ratio (A/Cr) >3–30 mg/mL, albuminuria A/Cr >30–100 mg/mmol and >100 mg/mmol was present in nine and five, respectively. Four had GFR <60 mL/min/1.73 m² and one was on maintenance dialysis.

The “control” group had a wide range of renal disease, mainly glomerular (lupus in 5, 2 each with diabetes, focal glomerulosclerosis, crescentic glomerulonephritis and vasculitis), and a variety of other diagnoses. As well there were three with acute tubular necrosis (ATN) and one normal individual.

Fresh midstream urine specimens from both groups were centrifuged and resuspended as 25 μL of the thus ×20 concentrated specimen. This 25 μL specimen was examined under polarised light in a counting chamber. The number of Maltese cross particles/μL was graded: None = 0, <100=1, >100 with no clumps =2 and >100 with clumps =3.

Three types of Maltese cross particles were found. Unfortunately, but perhaps educationally, subjects had been issued with starch-covered latex gloves to collect their urine specimens. Maltese cross particles easily identifiable as starch granules were seen in many specimens. Rounded, non-lamellated particles containing symmetrical Maltese crosses were seen in all but one albuminuric individual, both in Fabry and other renal diseases, but not in one patient with membranous nephropathy in remission (A/Cr 26 mg/mmol). They were not seen in the ATN patients or in the normal individual. These were thus typical urinary lipid particles (Figures 1–3).

Importantly a different type of Maltese cross particle, mostly spherical or non-spherical with protrusions of different size and shape, less frequently a structure with an internal spiral pattern always with an asymmetrical, truncated or atypical Maltese cross, was seen in 28/29 Fabry patients (Figure 5). No particles with this appearance were seen in the 21 “controls”. In the Fabry patients there was a direct relationship between the severity of albuminuria and the graded number of particles seen.

Figure 5:

Fabry disease: a lipid containing particle with protrusions (A) and a round particle with internal spiral pattern (C), both particles with asymmetrical Maltese crosses under polarised light (B, D).

(Courtesy of G.B. Fogazzi and G. Garigali).

To further explore the identity of the particles in the Fabry urine we used immuno- histochemistry to identify cells as podocytes using anti-human podocalyxin antibody (PHM5, Clone 18.29; Chemcon catalogue, Millipore, Billerica, MA, USA) and GL3 by an anti CD 77 antibody (Abcam, Cambridge, UK). By this technique we were able to identify GL3 in renal podocytes in patients Fabry urine [17]. We suggest that using the examination of urinary sediment with polarised light microscopy may be a useful way to detect Fabry disease particularly in situations where enzyme or genetic analysis is unavailable.