Histological diagnosis of renal osteodystrophy

Transcription

1 Kidney International, Vol. 56, Suppl. 73 (1999), pp. S-26 S-30 Histological diagnosis of renal osteodystrophy TONY FREEMONT Department of Pathology, University of Manchester, Manchester, United Kingdom Several other factors can affect the morphology of the processed bone sample. Polymerization of the mono- mer which, by the end of processing, replaces all the water in the tissue can, if uncontrolled, lead to a vigor- ous exothermic outcome with the potential to generate sufficient heat to boil the tissue. Sectioning the tissue using a mechanically-powered microtome inevitably pro- duces compressive loads on the tissue in one plane, re- sulting in micro-distortion of the section. Staining the sections in water and/or alcohol-soluble dyes subjects the tissue, or rather the resin in which it is embedded, to physico-chemical forces (or, worse still, removes the resin, which eliminates support and physical constraints). Mounting the section on a microscope slide and applying high loads to it flattens the section. While these are technological problems, knowing that they occur is important to the end-user who must make clinical decisions based on the data obtained from analy- Histological diagnosis of renal osteodystrophy. Histology and histomorphometry are the two primary methods of using bone biopsies to diagnose renal osteodystrophy. However, appropriate diagnoses can only be rendered when there is complete understanding of bone structure, of they way bone is processed for histology, and of the manner in which analyses of samples proceeds. In the future, modern applied molecular biology techniques will be added to these standards and redefine diagnoses. The pathologist has two tools with which to make histologic diagnoses on bone biopsies: subjective histologic examination and histomorphometry. Before anyone can meaningfully interpret data derived from histologic examination, they must understand the way in which these two tools are used, their shortcomings, and their advantages [1, 2]. To understand both requires knowledge of the bone itself, how it is treated during processing for histologic analysis, and the way in which subjective and objective data are derived from it. THE BONE BIOPSY The bone is obtained, preferably as a transiliac core using a specially designed trephine needle [3], and immersed into fixative. The most commonly used fixative is alcohol, which preserves tissue by dehydration. Alcohol is used in preference to water-based cross-linking fixatives such as formaldehyde because of its ability to penetrate through fatty marrow to the bone surfaces where bone cells are found, as well as the absence in alcohol of hydrogen ions, which might decalcify the tissue. The disadvantages of alcohol are that it is poor at preserving macromolecular epitopes and inhibiting nucleases, rendering it considerably inferior to other fixatives for subsequent tissue probing by immunohistochemistry (IHC) or in situ hybridization (ISH). For this reason we split the biopsy into two longitudinally, processing one part specifically for ISH and IHC and the other for morphometry. This has proved to be a very useful strategy [4 8] by the International Society of Nephrology Processing tissue for examination For histomorphometry the biopsy must be sectioned undecalcified, a double negative indicating that this is not the usual way for bone to be prepared for micros- copy. The tissue must be cut very thinly in order for light to pass through the slice or section, a necessity engendered by the workings of the light microscope. Cut and unsupported, the bone would tear away from the softer marrow, destroying the interface, which is the single most important area of the biopsy, as it is here that most bone cells act. Thin sectioning is usually achieved by embedding the biopsy in a polymerized plastic resin of a similar consistency to bone matrix. To adequately infiltrate the biopsy with resin necessitates a number of processing steps, during which the tissue is immersed in a series of chemicals, all of which have the effect of causing the biopsy to change size. Although this effect may be imperceptible to the naked eye, it causes alter- ations easily detected in the light microscope. These alterations have implications for histomorphometry, a technique that relies on taking and comparing measurements. Size changes continue to occur during the later stages of processing and sectioning. Technological problems affecting bone data S-26

2 Freemont: Histological diagnosis of renal osteodystrophy S-27 sis of the sections. Fortunately, within any one laboratory that holds to one protocol, the effects are relatively uniform and predictable, allowing valid comparison between data derived from different biopsies within that laboratory. Much of histological diagnosis (including histomorphometry) is comparative for just this reason, but all diagnostic parameters are eventually compared with controls ; the best laboratories will have their own adequate control populations processed in the same way as their diagnostic biopsies [9]. DATA HELD WITHIN A BIOPSY At the moment the fixative permeates the biopsy, all metabolic processes are suspended and the pathologist is at his/her leisure to examine the snapshot of tissue events that were occurring within the tissue at that moment. The main approach is to stain the tissue using tinctorial methods in such a way as to render components of the tissue/cells distinguishable from one another [10, 11]. Interpretation is then based on distinctive morphology and/or coloring of the important components of the tis- sue. Interpretative features can be detected either by eye or by trained machines. Thus, tinctorial staining dis- tinguishes mineralized bone from osteoid, active cells from relatively quiescent ones and sites of metal deposi- tion from otherwise normal mineral [12, 13]. Morphology distinguishes large multinucleated cells from mononuclear ones and smooth from irregular bone surfaces. Many more data are held within the tissue than can be visualized by simple staining and modern techniques, applied to tissue sections, have enabled the cell and mo- lecular biology of the bone to be investigated [14 16]. This is where the techniques of ISH, IHC, etc. come in to their own. The advantage of these in situ techniques over others is that molecules can be localized to specific cells, an essential when the tissue contains so many disparate cell types. As in situ techniques become more sophisticated [8], a greater amount of information can be extracted and interpretation of tissue events becomes more precise. These methods will revolutionize the understanding and treatment of bone disease. Tetracycline One of the most important resources for interpreting events within bone is tetracycline labeling [17]. The tetracycline family of drugs has some very important and interesting properties. The drug is incorporated into mineralizing bone and other calcified tissues, and in ultraviolet light it fluoresces green to yellow (depending upon the exact molecular shape). If given to a patient prior to biopsy, it can be seen as a yellow-green line when unstained tissue sections are viewed in ultraviolet light. This line can be used as a marker of mineralization. If two or more doses of tetracycline are given at specified Fig. 1. Transiliac bone biopsy. (a) Three-dimensional reconstruction of part of a transiliac bone biopsy using a high energy X-ray source. (b) Microscope section prepared from the biopsy. Toluidine blue staining, magnification 100. intervals, parallel lines of fluorescence can be seen separated by unlabeled bone. The separation between the lines is a measure of the rate of mineralization and, in steady bone deposition states, the rate of bone formation. From this simple expedient it is possible to take the biopsy away from being a single snapshot in time and turn it into a source of dynamic data. Two-dimensional representations The microscope slide is effectively a two-dimensional representation of the tissue. This introduces a number of complexities into interpretation. Most of these relate to the fact that bone relies for its physical properties on its three-dimensional structure; objective and subjective measurements must somehow reconcile the need to ex- tend two-dimensional images into three-dimensional functional units. This problem can be illustrated by ex- amining four aspects of bone biology: Trabecular connectivity. The strength of many bones (e.g., vertebral bodies) depends upon the ability of tra- beculae to act as supporting rods. This property is a function of connectivity, a three-dimensional parameter (Fig. 1a). The histopathologist sees the bone trabeculae as two-dimensional structures (Fig. 1b). It is possible to take series of two-dimensional sections and superimpose them, mentally or using computerized reconstruction programs, but it is never possible to exactly reproduce the structure in this way without time-consuming and complex geometrical reconstructions. Pathologists are always looking for other ways to represent the three- dimensional structure of bone. The image in Fig. 1a is generated by the high-energy X-ray emissions from a synchrotron, a relatively nondestructive and time-effi- cient method for demonstrating connectivity. Unfortunately, although Manchester University has an instrument on-site, few laboratories have this level of access.

3 S-28 Freemont: Histological diagnosis of renal osteodystrophy Fig. 2. Two-dimensional representation of a circle (center) could be part of a three-dimensional sphere (left) or a three-dimensional cylinder (right). Fig. 3. Diagram of 10 histologic sections through a piece of bone containing an osteoclast. Sections 3 9 include the levels through the osteoclast. Serial sections 6 8 all contain the osteoclast, whereas in step serial sections 1 and 6 the osteoclast is only present in 6. Spheres and cylinders. This problem is illustrated by Figure 2. Simply put, the question is whether a circle seen in two dimensions is part of a cylinder or a sphere. At first sight this seems to be a very esoteric point, but when two-dimensional images are converted to three- The thickness of the label is dependent upon the length dimensional data (as recommended by the American of time over which the tetracycline is given. Many labora- Society for Bone and Mineral Research [18]), the differ- tories request that each tetracycline label be given over ences between spherical and cylindrical geometry intro- 3 or 4 days, giving relatively thick lines. Modern enhanceduce the opportunity for considerable calculation errors. ment microscopy means that even small amounts of label Such problems usually affect measurements of the cor- can be visualized. Consequently, we ask that patients tex, so some laboratories do not routinely offer bone are given 15 mg of tetracycline per kilogram of body cell or matrix measures for cortical bone. The majority weight over 12 hours, decreasing the artefact that comes of bone tissue in the body is cortical, and many diseases from obliquely sectioned saucers of tetracycline [19]. (e.g., hyperparathyroidism) manifest most in the cortex. As a consequence, we offer cortical data and thus the Subjectivity nature of the geometry of the cortical bone units is of No matter how careful the observer or complex the particular importance. computer program, the nature of bone is such that sub- Big cells. In reconstructing three-dimensional struc- jective measures will always influence interpretation of tures, there is a temptation to use serial sections of bone. histologic sections of bone. For instance, all stains of Some structures are larger than the section is thick. The osteoid will also stain non-mineralized fibrous tissue the most notable example is the osteoclast. Osteoclasts are same color. In low-turnover bone states, osteoid seams few in number and very large. A commonly used parame- may be very thin and approach the thickness (thinness) ter in assessing bone is osteoclast number, derived by of the fibrous tissue that covers all resting bone surfaces. counting osteoclast profiles in tissue sections and relating Assessing just what is osteoid and what is fibrous tissue these to tissue volume or surface. If sections are selected can be very difficult under these circumstances, making that are too close together (Fig. 3) the same osteoclast extent of osteoid surfaces a difficult parameter to define. appears in several sections. While this would have no The same applies to distinguishing between osteoblasts effect if one were measuring the volume of the tissue and obliquely sectioned resting surface cells. The most occupied by osteoclast, when counting cells per unit area notoriously subjective area of bone histomorphometry or length of bone surface it introduces a multiplication is assessment of resorptive surfaces, which despite many area that could lead to misinterpretation, e.g., mistaking attempts has yet to be successfully automated. Some normal bone for bone in a high turnover state such as choose to use surface irregularity (in the absence of an hyperparathyroidism. For this reason we perform histo- osteoid surface) as the criterion, while others use erosion morphometry on step serial sections, i.e., sections through collagen lamellae assessed in polarized light taken every m. (Fig. 5.). There is no universal agreement as to how this Lines or saucers. In biopsies, thin structures such as parameter should be derived. tetracycline labels appear as lines (Fig. 4), and it is tempting to think of them as being effectively one-dimensional. Controls Tetracycline labels are not lines but are very thin, saucer- The subject of controls has been alluded to already. shaped deposits incorporated into bone by a sheet of Differential shape change induced by processing and osteoblasts. If the sectioning is not tangential to the edge sectioning will affect the size and shape of tissue section of the saucer, the line of the tetracycline is not sharp profiles, and this has influenced our decision to have in- but is thickened and indistinct. Clearly, the thinner the house controls [9]. There is considerable variation in tetracycline label, the less of a problem this becomes. bone parameters in normal individuals with sex and

4 Freemont: Histological diagnosis of renal osteodystrophy S-29 Fig. 4. Unstained histological section viewed in UV light. The tetracycline appears bright. Various artefacts caused by sectioning are seen in the biopsy, including effects due to oblique sectioning. (Original magnification 200). Fig. 5. Section of bone viewed in polarized light to show one way of identifying resorption surfaces. (Original magnification 250). age and so our normal values cover a wide range of ages and both sexes in 10 years cohorts. Even these data are limited as they only pertain to biopsies from the anterior iliac crest of Caucasians. Other bones, other sites in the iliac bone, and other ethnic groups will have different measures. The generation and interpretation of biopsies is therefore subjective, complex and unusually observer depen- dent. It can be overcome by using a laboratory with long- standing standardized technology, in-house controls and, unusually for scientific studies, only one data generator. This is not, however, the only problem confronting the pathologist. This laboratory receives between 100 and 150 bone biopsies from renal patients from 8 countries each year, and has done so for more than 30 years. It is becoming apparent that renal bone disease is changing with time [20, 21]. For example, adynamic bone was not a feature of our practice until 1985, yet it now accounts for 40% or more of the biopsies received [22 24]. In a similar vein, renal bone disease is different in different dialysis and transplantation centers around the world. While typical hyperparathyroid bone disease is seen in Mexico and Greece, it is seen increasingly rarely in England or Sweden; in the latter two countries, peritrabecular marrow fibrosis and increased osteoclasis, hallmarks of hyperparathyroid bone disease, are increasingly associ- ated with relatively poor osteoblast function, perhaps because of resistance to parathyroid hormone [23, 25, 26]. It is becoming increasingly difficult to shoe horn patients into typical disease categories, e.g., adynamic bone, hyperparathyroid bone disease, osteomalacia. Be- cause of these problems, we have adopted a different approach towards the analysis of bone biopsies in patients with renal dysfunction. All bone disease is a conse- quence of disturbed bone cell function. Effectively, there are 4 bone cells: osteoblast, osteoclast, resting surface cell and osteocyte. If adequate technology existed, it would be possible to define all bone diseases in terms of the function of these four cell types. Despite introduc- ing in situ molecular techniques (IHC, ISH) into our diagnostic armory, we are only scratching at the surface of understanding bone cell function and interactions. However, it is entirely possible to redefine bone disease in terms of the number and function of osteoblasts and osteoclasts based on current histomorphometric data. The only problem we find is selecting which parameters to use to define these 4 measures from among the 32 parameters we give to clinicians on every biopsy and the 21 that we measure or derive but don t release with each report. However, all the data indicate that with the development of further pharmaceutical and alternative therapeutic strategies, it might soon be possible to influ- ence the behavior and/or number of individual cell types. The prospect that biopsy-derived data could be the basis for treatment tailored to the needs of individual patients is an exciting challenge and opportunity. FUTURE POSSIBILITIES If the trends we are seeing are recognized in other countries, the need we have encountered to redefine bone disease in terms of differential cell function/number will lead to a reclassification of bone disease. This will be facilitated by modern laboratory techniques that en- able the molecular biology of bone disease to be explored in tissue, and hence within single cell types in their natu- ral physical and chemical environment. This holds out considerable hope for new, highly-targeted therapies for the management of bone disease in patients with renal dysfunction.

5 S-30 Freemont: Histological diagnosis of renal osteodystrophy REFERENCES receptor and IL-6 nuclear factor gene expression in Paget s disease. J Bone Min Res 9:75 80, Malluche HH: Renal bone disease an ongoing challenge to the 16. Langub MC Jr, Koszewski NJ, Turner HV, Monier-Faugere nephrologist. Clin Nephrol 44 (Suppl 1):S38 S41, 1995 MC, Geng Z, Malluche HH: Bone resorption and mrna expres- 2. Malluche HH, Monier-Faugere MC: The role of bone biopsy sion of IL-6 and IL-6 receptor in patients with renal osteodystroin the management of patients with renal osteodystrophy. JAm phy. Kidney Int 50: , 1996 Soc Nephrol 4: , Heaf JG, Podenphant J, Gammelgaard B: The reliability and 3. Lalor B, Freemont AJ, Carlile S: An improved transilial bone representativity of non-dynamic bone histomorphometry in uremic biopsy drill for quantitative histomorphometry. Bone 7: , osteodystrophy. Scand J Urol Nephrol 27: , Parfitt AM, Drezner MK, Glorieux FH, Kanis JA, Malluche 4. Hoyland JA, Freemont AJ: Investigation of a quantitative poststandardization H, Meunier PJ, Ott SM, Recker RR: Bone histomorphometry: hybridization signal amplification system for mrna-oligodeoxyribonucleotide of nomenclature, symbols, and units. Report of the in situ hybridization. J Pathol 164:51 58, 1991 ASBMR Histomorphometry Nomenclature Committee. J Bone 5. Walsh L, Freemont AJ, Hoyland JA: The effect of tissue decalci- Miner Res 2: , 1987 fication on mrna retention within bone for in situ hybridisation 19. Hutchinson AJ, Freemont AJ, Boulton HF, Gokal R: Low studies. Int J Exp Pathol. 74: , 1993 calcium dialysis fluid and oral calcium carbonate in CAPD. A 6. Andrew JG, Hoyland JA, Marsh D: Insulinlike growth factor method of controlling hyperphosphataemia while minimizing alugene expression in human fracture callus. Calcif Tissue Int. 53:97 minium exposure and hypercalcaemia. Nephrol Dial Transplant 102, : , Braidman IP, Davenport LK, Carter H, Selby PL, Mawer EB, 20. Monier-Faugere MC, Malluche HH: Trends in renal osteodys- Freemont AJ: Preliminary in situ identification of estrogen target trophy: a survey from 1983 to 1995 in a total of 2248 patients. Nephrol Dial Transplant 11 (Suppl. 3): , 1996 cells in bone. J Bone Min Res 10:74 80, Sherrard DJ, Hercz G, Pei Y, Maloney NA, Greenwood C, 8. Hoyland JA, Mee AP, Baird P, Braidman IP, Mawer EB, Free- Manuel A, Saiphoo C, Fenton SS, Segre GV: The spectrum of mont AJ: Demonstration of estrogen receptor mrna in bone bone disease in end-stage renal failure: An evolving disorder. Kidusing in situ reverse transcriptase polymerase chain reaction. Bone ney Int 43: , :87 92, Hutchinson AJ, Whitehouse RW, Boulton HF, Adams JE, 9. Rehman MTA, Hoyland JA, Denton J, Freemont AJ: Age re- Mawer EB, Freemont AJ, Gokal R: Correlation of bone histology lated histomorphometric changes in bone in normal British men with parathyroid hormone, vitamin D3 and radiology in end-stage and women. J Clin Pathol 47: , 1994 renal disease. Kidney Int 44: , Freemont AJ, Denton J: Demonstration of aluminium in bone 23. Cueto-Manzano AM, Konel S, Hutchison AJ, Crowley V, using different staining techniques and spectrophophotometry. J France MW, Freemont AJ, Adams JE, Mawer B, Gokal R: Clin Pathol 42: , 1989 Bone loss in long-term renal transplantation: histopathology and 11. Hutchinson AJ, Lumb GA, Gokal R: Renal osteodystrophy in densitometry analysis. Kidney Int 55: , 1999 CAPD. Adv Perit Dial 7: , Malluche HH, Monier-Faugere MC: Risk of adynamic bone 12. Byers RJ, Denton J, Hoyland JA, Freemont AJ: Differential disease in dialyzed patients. Kidney Int 42 (Suppl 38): S62 S67, patterns of osteoblast dysfunction in trabecular bone in patients 1992 with established osteoporosis. J Clin Pathol 50: , Hutchinson AJ, Whitehouse RW, Freemont AJ, Adams JE, 13. Nicholas JCB, Dawes PT, Davies SJ, Freemont AJ: Persisting Mawer EB, Gokal R: Histological, radiological and biochemical aluminium-related bone disease after cadaveric renal transplanta- features of the adynamic bone lesion in continuous ambulatory tion. Nephrol Dial Transplant 14: , 1999 peritoneal dialysis patients. Am J Nephrol 14:19 29, Andrew JG, Hoyland JA, Andrew SM, Marsh D, Freemont 26. Gerakis A, Hutchinson AJ, Apostolou T, Freemont AJ, Billis AJ: Demonstration of TGF- 1 in normally healing fracture callus A: Biochemical markers for non-invasive diagnosis of hyperparaby in situ hybridisation. Calcif Tissue Int 52:74 78, 1993 thyroid bone disease and adynamic bone in patients on haemodial- 15. Hoyland JA, Freemont AJ, Sharpe P: Interleukin 6 (IL-6), IL-6 ysis. Nephrol Dial Transplant 11: , 1996