Nephrotoxicity Testing In Vitro

The Report and Recommendations of ECVAM Workshop 101,2

Reprinted with minor amendments from ATLA 23, 713-727

Gabrielle M. Hawksworth3, Peter H. Bach4, J. Fred Nagelkerke5, Wolfgang Dekant6, Jacques E. Diezi7, Ernest Harpur8, Edward A. Lock9, Caroline MacDonald10, Jean-Paul Morin11, Walter Pfaller12, Fons A.J.J.L. Rutten13, Michael P. Ryan14, Hervé J. Toutain15 and Andrea Trevisan16
3Departments of Medicine & Therapeutics and Biomedical Sciences, Polwarth Building, University of Aberdeen, Aberdeen AB9 2ZD, UK; 4Faculty of Science, University of East London, Romford Road, London E15 4LZ, UK; 5LADCR, Division of Toxicology, Center for Bio-Pharmaceutical Sciences, Leiden University, 2300 RA Leiden, The Netherlands; 6Universitat Wurzburg, Versbacherstrasse 9, 97074 Wurzburg, Germany; 7Institut de Pharmacologie et de Toxicologie de l'Universite, rue de Bugnon 27, 1005 Lausanne, Switzerland; 8Sanofi-Winthrop Research Centre, Willowburn Avenue, Alnwick, Northumberland NE66 2JH, UK; 9ZENECA Central Toxicology Laboratory, Alderley Park, Macclesfield, Cheshire SK10 4TJ, UK; 10Department of Biological Sciences, University of Paisley, High Street, Paisley PA1 2BE, UK; 11INSERM U295, Uer Medicine Pharmacie, Université de Rouen, 76803 Saint Etienne Rouvray, France; 12Institut für Physiologie und Balneologie, Universität Innsbruck, Fritz-Preglstrasse 3, 6010 Innsbruck, Austria; 13TNO Nutrition and Food Research Institute, Division of Toxicology, 3700 AJ Zeist, The Netherlands; 14Department of Pharmacology, University College Dublin, Fosters Avenue, Blackrock, Dublin, Ireland; 15Rhone-Poulenc Rôrer, Centre de Recherche de Vitry, Quai Jules Guisde, 94405 Vitry Cedex, France; 16Instituto di Medicina del Lavoro, Universita di Padova, Via Facciolati 71, 35127 Padova, Italy

1ECVAM - The European Centre for the Validation of Alternative Methods. 2This document represents the agreed report of the participants as individual scientists.

Address for correspondence: Dr G.M. Hawksworth, Department of Medicine and Therapeutics, Polwarth Building, University of Aberdeen, Foresterhill, Aberdeen AB9 2ZD, UK

Address for reprints: ECVAM, TP 580, JRC Environment Institute, 21020 Ispra (VA), Italy

Preface

This is the report of the tenth of a series of workshops organised by the European Centre for the Validation of Alternative Methods (ECVAM). ECVAM's main goal, as defined in 1993 by its Scientific Advisory Committee, is to promote the scientific and regulatory acceptance of alternative methods which are of importance to the biosciences and which reduce, refine or replace the use of laboratory animals. One of the first priorities set by ECVAM was the implementation of procedures which would enable it to become better informed about the state-of-the-art of non-animal test development and validation, and the potential for the possible incorporation of alternative tests into regulatory procedures. It was decided that this would be best achieved by the organisation of ECVAM workshops on specific topics, at which small groups of invited experts would review the current status of various types of in vitro tests and their potential uses, and make recommendations about the best ways forward (1).

The workshop on Nephrotoxicity Testing In Vitro was held in Angera, Italy, on 16-20 May 1994, under the co-chairmanship of Gabrielle Hawksworth, Peter Bach, and Fred Nagelkerke. The participants comprised scientists working in both academia and industry, and included representatives from five of the six groups which participated in the recent European Commission (DGXII)-funded BRIDGE (Biotechnology Research for Innovation, Development and Growth in Europe, 1990-1994) nephrotoxicity project. This report describes the systems currently employed for nephrotoxicity testing, the methods used for the determination of various endpoints, and the applications of these methods in industry; it also outlines future developments in the field. As a result of the discussions which took place at the workshop, a number of recommendations have been made concerning the use of in vitro systems for nephrotoxicity testing, which are given at the end of this report.

Systems in Current Use

The in vitro systems currently employed for nephrotoxicity studies were reviewed during the workshop; the advantages and disadvantages of each system for undertaking toxicological studies are summarised below.

Isolated Perfused Kidney

The isolated, cell-free, perfused (rat) kidney [IP(R)K] is the most appropriate system for studying potentially nephrotoxic xenobiotics when tubulo-vascular integrity is required (2). This system is not influenced by extrarenal regulatory systems (for example, nervous, hormonal, and blood-borne factors). The IPK enables precise control of the concentrations of compounds which are being studied. However, it is not a system to be used for routine studies.

The disadvantage, compared with studies of the kidney in situ, is that renal function is maintained for only a relatively short period (up to four hours), although the limited lifetime of this preparation can be improved considerably by adding oxyphoretic compounds to the perfusate (3). The marked decrease in concentrating ability leads to an elevated fractional excretion of water. The perfusion system also leads to higher renal vascular flow (RPF) and lower glomerular filtration rates (GFR), as well as to decreased filtration fractions (FF) and fractional sodium excretion (FENa), relative to the in vivo situation.

Three different modes of perfusion are available: a) single pass perfusion; b) recirculation of the perfusate (which minimises the costs incurred due to adding albumin, etc. to the perfusate); and c) recirculation and dialysis of the perfusate. Each of these requires optimal oxygenation of the perfusate. The methodology has been described by Bowman (4), Maack (5), and Bekersky (6). Minimum criteria which should be met when using the IPRK have been outlined in Table I. Experiments should be carried out using sterile materials and solutions, since this will result in longer term viability of the preparation. It should be noted that data obtained using the IP(R)K may differ from those obtained from kidney perfusion studies in situ.

Table I: Conditions for Using the Isolated, Perfused Rat Kidney

Parameter Value
Perfusion pressure(PA) 80-110 mmHg
Perfusate flow(RPF) 20-35ml/min
Fractional sodiumexcretion (FENa) < 7%
Glomerular filtrationrate (GFR) > 0.5ml/g kidney

Isolated Perfused Nephron

The isolated perfused nephron is not suitable for routine nephrotoxicity studies. However, investigations conducted with the non-perfused nephron have provided valuable data on enzyme localisation and receptor distribution, which can be used for confirming the site of origin of isolated cell systems (7).

Renal Cortical Slices

Renal cortical slices have been used extensively for the study of renal transport and toxicity, typically for no more than a two-hour period (8-11). More recently, the use of precision-cut renal cortical slices has become fashionable. They are easy to produce, and differences in the preparation of slices between investigators are minimised by using commercially available tissue slicers. Precision-cut renal cortical slices have been used for periods of up to 24 hours (12-14). The advantages and disadvantages of renal cortical slices are summarised in Table II.

Table II: Advantages and Disadvantages of Renal Cortical Slices

Advantages Disadvantages
  • Simple technique
  • Multicellular system
  • No enzymatic tissue digestion
  • Three-dimensional structure
  • Cell-cell contacts maintained
  • Site-specific effects can be studied
  • Several functional parameters can be assessed
  • Renal-specific parameters are maintained
  • Rapid and simple species comparisons can be undertaken
  • Limited lifetime
  • Morphological, functional, and biochemical heterogeneity
  • Lack of reproducibility of nephron segments
  • Lumens may be collapsed (hence only passive transportof p-amino-hippuric acid)
  • Surface of slice represents damaged area

Although this system still needs to be optimised, it shows considerable promise. The following aspects need to be studied further: optimisation of the culture conditions (for example, medium, temperature, and pO2), preparation time, slice thickness, regional distribution (cortical slices should comprise all of the nephrons, from the surface to the outer medulla), definition of suitable criteria for assessing viability and biotransformation capacity and for conducting species and sex comparisons. Successful cryopreservation of renal slices would enable more efficient use of human tissue and permit the transfer of material between research groups (15).

This relatively simple technique offers the opportunity to study both renal function and toxicity in different regions of the kidney (i.e. cortex, medulla or papilla), in an in vitro system that exhibits many of the characteristics seen in vivo.

Isolated Glomeruli and Proximal Tubular Fragments

Freshly isolated fragments of renal tissue have been used for assessing the acute effects of chemicals on glomeruli and proximal tubules, including comparisons of superficial and deep glomeruli. The isolation is based on forcing small blocks of tissue (either untreated or subjected previously to enzymatic digestion), in oxygenated buffer, through a series of stainless steel sieves. The glomeruli are resistant to shearing forces and are harvested on a 63 µm (rat) or a 106 µm (human) sieve. The proximal tubules are collected on a 50µm (rat) or a 75 µm (human) sieve. The purity of the fragments is high (> 90% for both glomeruli and proximal tubules), and can be improved by subjecting the preparation to Percoll isopycnic density gradient centrifugation (16). Alternatively, iron oxide particles can be infused into the intact kidney. These lodge in the glomeruli, which facilitates the collection of higher purity glomeruli by magnetic separation (17), and also results in the production of an improved preparation of proximal tubules due to the exclusion of glomeruli. Renal fragments can also be prepared by microdissection.

The glomeruli or tubular fragments can be used to measure a whole range of non-specific, but sensitive, endpoints (see later), as well as a number of processes specific to the cell type of interest. For glomeruli, this would include metabolism, and the synthesis of proteoglycan, collagen, and fibronectin; with proximal tubules, metabolism and transport studies can be undertaken. Studies have been undertaken with renal fragments isolated from rat, rabbit, pig, dog and human kidneys. To isolate the fragments, the same basic technique as for isolating cells from these fragments is used (see the section below).

The advantage of using renal fragments is that they retain the architecture of the anatomical regions of the kidney from which they are derived. They offer a simple, rapid, inexpensive, and flexible method for measuring various endpoints in glomeruli and proximal tubules, both of which can be isolated from the same kidney. Renal fragments can also be used to assess alterations in biochemical functions at various times after injury, and may provide useful information about mechanisms of toxicity (18). The limitations of the method are that the surfaces of the renal fragments represent damaged areas, and thus the fragments are only viable for four to six hours at the most; also, the lumen of the tubule may be collapsed, which would preclude their use for luminal transport studies. The isolation of fragments from kidneys with well developed fibrous networks (i.e. from older animals or humans) may be less successful, unless enzymatic digestion is used during the isolation procedure. However, it is possible to extend the lifetime of these preparations by using sterile conditions and optimised medium, and by the addition of antioxidants during the isolation procedure.

Isolated Proximal and Distal Tubular Cells

Cells are isolated by two-stage collagenase perfusion, as originally described by Ormstadt et al. (19). The animal is anaesthetised and, after laparotomy, the aorta is cannulated and the kidneys are perfused with a buffer solution containing collagenase. After excision of the kidneys, the capsule is removed and the tissue is dispersed in buffer. The suspension is then filtered through nylon gauze or Endecott sieves, and the filtrate is subjected to Percoll density gradient centrifugation. Rat proximal tubular cells band at a density of 1.02-1.05g/ml; rat distal tubular cells band at a density of 1.08-1.12 g/ml. Alternatively, for the isolation of cells from human or rat kidneys, cortical tissue can be minced coarsely prior to two-stage collagenase perfusion and Percoll density gradient centrifugation (20, 21). It is recommended that about 10g human kidney cortex, or two to four rat kidneys, are used. The yield of cells from rat kidney is lower using the latter method, but the procedure is simpler and the functional capacity of the cells is not compromised. Identification of the cells is carried out by determining their enzyme distribution, receptor responsiveness and transport capabilities (22). Tubular cells can also be isolated without the use of digestive enzymes, for example, by free flow electrophoresis (23).

These preparations have been used extensively for acute nephrotoxicity studies. The cells retain most of the characteristics seen in vivo. They can, therefore, be used to study mechanisms of toxicity at the cellular level, and to design strategies for reducing/protecting against cell toxicity (i.e. for cytoprotection). Relevant in vitro-in vivo extrapolations can also be undertaken, and comparisons can be made between species. However, cell suspensions have a limited lifetime (up to six hours), and primary cell cultures are needed for studies over a longer period. The use of primary cell cultures also enables the effects of lower, and possibly more relevant, concentrations of chemicals to be investigated.

Primary Cell Cultures

A number of different procedures are used to obtain primary cell cultures. Microdissection techniques have been used to establish explant cultures for all nephron segments. The origin of the cell type is well defined, but microdissection usually results in a low yield, and growth factors and serum-containing medium are needed. The proximal tubular cells obtained are responsive to parathyroid hormone, but not to arginine, vasopressin or calcitonin. They show low activities of alkaline phosphatase, y-glutamyltranspeptidase and Na+/K+-ATPase. The technique can be used to isolate human proximal convoluted tubular cells and proximal straight tubular cells separately (24).

Two immunologically based techniques have been described. Immunodissection has been used for isolating rat and human proximal tubular cells, using a monoclonal antibody against microvillus membrane proteins of rat renal cortical cells. The cells show sodium-dependent glucose transport and Na+/H+ exchange. The second immunological method involves flow cytometry, and has been used to isolate human proximal and distal tubular cells (from the same sample) by employing different monoclonal antibodies (25). The main advantage of this technique is the very high purity of the cell preparation obtained, although the yield is low and it is a time-consuming procedure, taking about 20 hours. Nevertheless, it is anticipated that, with the availability of more antibodies and with further technical developments, the full potential of this technique should be realised.

Higher cell yield is obtained using enzymatic digestion or mechanical dissociation followed by density gradient centrifugation (see the previous section). This procedure can be used to obtain isolated cells or three tofour cell fragments, which have a higher seeding efficiency than single cells (16). It results in cell preparations which are enriched in proximal tubular enzymes (for example, alkaline phosphatase) relative to distal tubular enzymes (for example, hexokinase). The method can incorporate glomerular separation by using iron oxides, but this is not essential (17). Mechanical dissociation of cortical tissue has mainly been used to isolate rabbit and rat proximal tubular cells (26). This isolation technique, which was originally described by Brendel & Meezan (27), yields preparations with higher intracellular ATP levels, respiration rates and transport activities than those for preparations isolated by proteolytic digestion and isopycnic density gradient centrifugation (28). Care should be taken to ensure oxygen saturation of the solutions during the isolation procedure. Good seeding efficiencies are obtained on uncoated plastic. Growth initiation can be improved by short, mild treatment with collagenase (5 mins; 0.125 mg/ml). When using small-well plates, it is important to check the seeding efficiency and cell growth. The proliferation of rabbit proximal tubular cells in primary culture decreases as the size of the wells decreases. Seeding densities in the range 1 x 105 - 2 x 105 cells/cm2 are used for human and rat proximal tubular cells.

The procedure used for isolating proximal tubular cells may be crucially important for subsequent cell growth requirements. Constant hyperoxygenation of cells should be employed to obtain optimal cell seeding efficiency. The most widely used culture medium is 1:1 Dulbecco's Modified Eagle's medium/Ham's F12. For optimal attachment after Percoll separation, fetal bovine serum, collagen-coated plastic, or both of these, are needed. The medium usually contains hydrocortisone/dexamethasone and insulin. Following mechanical dissociation, the cells have much less stringent growth requirements, and the use of fetal calf serum (FCS) for cell attachment should be avoided because of its detrimental effects on cell differentiation. With rabbit proximal tubular cells, the use of glucose-free and insulin-free medium may improve cell differentiation, and the maintenance of carbohydrate metabolism (by preventing loss of the gluconeogenic pathway), brush border enzyme activities and cellular architecture.

The medium should contain only low concentrations of antibiotics (25 g/ml for the first 48 hours) and no antifungals. Using porous membranes, culture of cells on Type IV collagen is the most appropriate method for maintaining both organic anion and cation transport (29). Culture on porous membranes is essential to study vectorial transport and results in better maintenance of the brush border in proximal tubular cells. It also enables measurement of the confluency and intactness of the renal epithelial cell barrier. Coating the porous membranes with 'Matrigel' or other substrata also results in better differentiation, and the cellular morphology more closely resembles that observed in vivo. It has been shown that the culture of Madin Darby canine kidney (MDCK) cells on Type I collagen results in dedifferentiation, with the induction of a fusiform shape and the loss of apical-basal polarity (30). More data are needed on the maintenance of enzyme activities, such as those associated with glutathione-S-transferases and cysteine conjugate C-S lyases, which are strongly implicated in the nephrotoxicity of several xenobiotics. Proximal tubular cell cultures from rat, rabbit, dog, and humans have been described, but there are no data on cultures derived from either the marmoset or the mini-pig. In addition, better methods are needed for the culture of rat proximal tubular cells. It is difficult to obtain confluent cultures of rat proximal tubular cells in the absence of serum and growth factors, but these additives are detrimental to cell differentiation.

Isolated human proximal tubular cells can be cryopreserved in FCS containing 10% dimethyl sulphoxide, and they show good recovery and viability post-thawing. However, rabbit proximal tubular cells do not recover and grow in culture when cryopreserved using similar conditions.

Separation by Percoll density gradient centrifugation can also be used to obtain a cell population enriched with distal tubular cells (see the previous section), but these cells have not been purified to the same extent as proximal tubular cells (25). The criteria employed for assessing distal tubular function are the transport characteristics of the cells.

Cells from the medullary thick ascending loop of Henlé (mTALH) have been separated by using centrifugal elutriation or immunological methods; they are characterised by the presence of the Na+/K+-cotransporter (31).

Human mesangial and epithelial cells from the glomerulus are isolated following dissociation of the tissue and collection of the glomeruli on a 106µm sieve. The glomeruli are cultured, and epithelial cells can be seen growing out of the glomeruli after 7-10 days. After two to three weeks, the mesangial cells overgrow the glomerular epithelial cells (32). The disadvantage of this method is that the cultured cells cannot be compared with freshly isolated cells to ensure that dedifferentiation has not occurred. Morphological and immunohistochemical, rather than functional, markers are used for identification purposes. These cells have not been widely used for toxicity studies, although the effect of cadmium on proteoglycan sulphation and excretion by cultured rat mesangial cells has been investigated (33).

Medullary interstitial cells have been isolated from rats, rabbits, dogs, and, more recently, from humans. For species other than man, cultures have been established from cells derived from subcutaneous autotransplants or isogeneic transplants, or by selective growth of cells in hypertonic medium (34). Human medullary interstitial cells have been isolated from primary explants of papillae by selective overgrowth of the interstitial cells. They are characterised morphologically and histochemically, and by the determination of prostaglandin E2 production and angiotensin II responsiveness (35). These cells need to be further characterised and compared with readily available cell lines, such as 3T3 cells, to assess their usefulness for toxicity studies.

Renal Cell Lines

Cell lines derived from the kidneys of various species are commercially available (Table III). It should be noted that differences in cell lines obtained from the same source have been observed. The cell lines used most in toxicological studies are those which exhibit in vitro some of the properties of the proximal tubular epithelium (PTE) in vivo, such as LLC-PK1, LLC-RK1, and OK cells. In culture, these cells grow to form a confluent monolayer, and they exhibit some degree of differentiation (for example, they show the morphological asymmetry of PTE cells which is observed in vivo). They have been shown to exhibit apical-basolateral polarisation, with brush border microvilli and basolateral infoldings being present. The expression of brush border microvilli is greater with LLC-PK1 cells than with the LLC-RK1 and OK cell lines, and is similar to that observed in the PTE in vivo. OK cells can be used to study parathyroid hormone-dependent sodium and phosphate transport. The expression of differentiated functions specific to the PTE varies with the particular cell line, and has been reviewed in detail by Gstraunthaler (36). The level of expression may differ according to the passage number, and with the composition of the culture medium, in particular in the presence of FCS or in hormonally defined medium (37).

Table III: Some Readily Available Renal Epithelial Cell Lines

Cell Line Species of Origin Nephronal Origin Catalogue No.
LLC-PK1 Pig Proximal Tubule CRL 1392a
LLC-RK1 Rabbit Proximal Tubule CCL 106a
OK Opossum Proximal Tubule CRL 6551a
JTC-12 Monkey Proximal Tubule RCB 456b
MDCK Dog Distal Tubule/Collecting Duct CCL 34a
MDBK Cow Cortex CCL 22a

aAmerican Tissue Culture Collection
bJapanese Tissue Culture Association

Transepithelial transport may be evaluated if the cells are grown on porous membranes. Unfortunately, the permanent cell lines which have been characterised with respect to their transport properties lack an organic anion transporter; the only exception to this is the OK cell line, which has been shown to accumulate p-aminohippurate.

The main advantages of renal cell lines are that laboratory animals are not needed to obtain tissues, their ease of handling, and the speed with which experiments can be conducted. However, none of the cell lines available show complete expression of those functions expected of proximal tubular cells. They often lack both Phase I and Phase II metabolising enzyme systems, and receptor-linked signalling systems. Another major disadvantage is the lack of useful cell lines which are representative of the different parts of the nephron, and which have been derived from the key species used in toxicological studies, such as the rat and dog, and from primates and humans. More information is required on the metabolic capabilities of existing cell lines, since biotransformation is often an important determinant of the mechanism of chemical-induced nephrotoxicity.

Hybrid Cells

Fusion of primary cells with established cell lines derived from the same region of the nephron may yield new cell lines which better retain some in vivo characteristics. Classically, mouse B lymphocytes can be immortalised by fusion with a murine myeloma cell line; the hybrid cells retain the expression of antibody (38). However, with other cell types, the hybrid cells do not always show an immortal phenotype, nor is expression of differentiated phenotypes always observed. Furthermore, hybrid cells show chromosomal segregation and,therefore, do not have a stable phenotype.

Immortalised Cells

The approach to cell immortalisation is to introduce oncogenes, for example, SV40 T antigen, myc or adenovirus E1A, into cells by gene transfer, in order to overcome cellular growth mechanisms. These oncogenes act by binding to normal cellular proteins which are encoded by tumour suppressor genes, thereby permitting progression through the cell cycle. One approach is to introduce the gene into the embryo and then to isolate cell lines from the transgenic animal. This works well for the mouse, but is not so useful for other species. Alternatively, the gene can be added to primary cultures of cells, for example, by calcium phosphate precipitation. The oncogene can be expressed constitutively or can be placed under the control of an inducible promoter (for example, heavy metal induction is used for the metallothionein promoter). The latter approach provides the opportunity to switch between cell growth and expression of the differentiated function (39, 40).

Immortalised cell lines have been isolated from the proximal tubule from transgenic mice (41), and by the transfection of rabbit mTALH (42, 43), rabbit proximal tubule (43), rat kidney epithelium (44), and human embryonic kidney (45, 46) cells. Cell lines have also been isolated by SV40 infection (39, 40). Many of these express tissue-specific characteristics, such as the expression of Tamm-Horsfall glycoprotein and ion transport systems. They have not, to date, been used for toxicity studies.

Future developments should include the characterisation of existing cell lines with respect to their xenobiotic metabolising activities, and the isolation of more cell lines which express the oncogene in a regulatable manner. Isolation of cell lines from different species would enable comparative data on interspecies differences in metabolism to be generated.

Genetically Engineered Cells

Cell lines which have been transfected with cloned genes, for example, with genes for transport protein receptors or xenobiotic metabolising enzymes, provide a powerful tool for studying mechanisms of action, since it is possible to investigate a particular system in isolation (47). They are less likely to be useful for in vitro toxicology studies, because not all of the systems involved in bioactivation or detoxification are present.

Subcellular Fractions

Fractionated organelles and membranes prepared from the kidneys in general, or from defined cell types, can be used for specific cell-free investigations. These in vitro systems, which are usually prepared by differential centrifugation, include nuclei, mitochondria (48), lysosomes (49), and membrane vesicles (50). The purity of the preparations is confirmed by marker enzyme assays, and by transmission and scanning electron microscopy. Lysosomal membrane integrity is evaluated by measuring enzyme latency (49), and mitochondrial function is determined by respiratory control assessment (48). Brush border and basolateral vesicles can be prepared from the same proximal tubular cells, and they can be used to study the membrane-specific transport of xenobiotics and the effects of chemicals on the transport of endogenous compounds.

Databases

The use of databases for predicting the potential nephrotoxic effects of chemicals is an area which needs to be developed, since it could reduce the duplication of studies and decrease the use of laboratory animals. The major problems with this approach are the reliability of, and the differing conditions used for, many of the in vivo and in vitro studies which have been reported in the literature. It is recommended that databases should incorporate data on the toxic response patterns of known nephrotoxins. Information submitted by industry (unattributable, if possible) would be very valuable, particularly where in vivo nephrotoxicity data are available. Otherwise it may be necessary to conduct in vivo studies to generate these data.

Endpoints of Toxicity

Morphological

Light microscopy is widely used for assessing cellular changes. The perfused kidney, renal tissue slices, and tissue/cellular fragments can be fixed and processed by routine histopathology techniques, and can be assessed subjectively or quantitatively. The use of histochemical and immunohistochemical techniques can provide additional information which is relevant to both screening and mechanistic studies. The endpoints that are commonly used include lethal (necrosis) and sub-lethal cell injury, and the loss or accumulation of those cellular structures, macromolecules and/or activities which can be detected histochemically.

The use of light microscopy is especially appropriate for assessing cells in culture by phase-contrast or fluorescence microscopy. These techniques are most commonly used to monitor, qualitatively, cell growth and dynamic changes; in combination with video microscopy, these changes can be quantified. Useful endpoints include cell counting, determination of the differentiated structure (domes), and the maintenance of specific functions. The use of confocal microscopy and appropriate fluorescent probes (for example, to study gap junctions) provides quantitative information on the distribution and dynamics of subcellular structures, such as lysosomes, the cytoskeleton and membrane proteins. Time-lapse video microsopy, image analysis, and spectroscopic measurements at the cellular level (spectrofluorimetry, single photon counting, and confocal microscopy) have been used to facilitate documentation of the sequence of cellular changes (such as alterations in ion distribution, volume and intracellular transport processes) which occur following exposure to chemicals, and also during any subsequent recovery of the cells (51). Subtle subcellular changes can be assessed qualitatively by electron microscopy.

Microscopy thus forms the bridge to in vivo toxicology, especially for the more complex in vitro systems. The application of new morphological techniques combines assessment of both biochemical and functional data at a cellular level. The disadvantage is that many of the techniques alter or kill the cells, so that these data represent single, discontinuous endpoints. There is also the risk of introducing bias if subjective assessments are made. It is recommended that quantitative image analysis should be used more widely, especially in combination with histochemical methods.

Functional

Various functional endpoints, and their applicabilities in the in vitro systems described earlier, are summarised in Table IV.

Table IV: Functional Endpoints Used in In Vitro Nephrotoxicity Assays

Endpoints Applicability Comments
Cellviability
  • Dye exclusion/retention
  • Enzyme/ion leakage
  • Renal cortical slices
  • Renal fragments
  • Isolated cells (in suspension)
  • Primary cell cultures
  • Cell lines
  • Applicable in intact cell systems and slices (forexample K+ content)
  • Enzyme leakage provides information on regiospecificity
  • NB:MTT assay is not specific for mitochondrial damage
Synthesisof macromolecules
  • Protein
  • DNA/RNA
  • Renal cortical slices
  • Renal fragments
  • Isolated cells
  • Primary cell cultures
  • Cell lines
 Check uptake of precursors
  • Matrix elements
  • May be useful in glomerular/ interstitial cell cultures
Rateof proliferation
(Clonogenic assays)
  • Cell lines
  • Total protein measurements (usually reflect cell number)
Carrier-mediatedtransport
  • Glucose
  • Systems containing proximal tubule cells
  • Can be quantified by uptake or transport
  • Organic Ions
  • Isolated tubules (transepithelial transport)
  • Cultures on porous membranes
  • Uptake applicable to all transepithelial systems
  • Check effects of specific inhibitors
  • Check paracellular transport
  • Not for routine use
  • Inorganic ions
  • Low molecular weight proteins
  • Not widely used
Endocytosis of labelled proteins or carbohydrates
  • Cultured cells of proximal tubular origin
  • Usually use radiolabel or horse radish peroxidase
  • Easy in vivo/in vitro comparison
Barrierfunction
  • All epithelial systems, intact epithelia or confluentcultures on porous membranes
  • Assessment of diffusion of extracellular markers,or electrophysiological measurements

Biochemical

The scheme outlined in Figure 1 illustrates how a chemical may interact with a renal cell to result in cellular injury and/or repair.

Figure 1: Biochemical Endpoints in In Vitro Nephrotoxicity Assays

Studies with the IPK can provide valuable information on the intrarenal handling of xenobiotics and their metabolites, and they may also provide information on selected endpoints, such as interactions with macromolecules, and effects on ATP synthesis and glutathione depletion (Figure 1). The IPK is the only system in which the integrated function of the whole organ can be studied, but it is not applicable for routine studies.

Well defined in vitro renal cell systems are ideal for studying the potential interactions of a chemical with cellular structures and functions. Initial studies should explore the time- and dose-response relationships. Further work should then be conducted at relevant, sublethal, doses. The concentrations used should be relevant to those delivered to the kidney in vivo. It is possible to elucidate the mechanisms of toxicity of chemicals by studying the biochemical cascade of events outlined in Figure 1, and to correlate the early effects of the interaction of a chemical (for example, the formation of protein adducts) with late endpoints relevant to observed changes in cell function. Flow cytometry is a promising new approach in which a number of biochemical parameters, including second messenger systems and [Ca]i and [pH]i, can be studied simultaneously with cell cycle events and DNA cleavage. Mechanisms of cell death, including apoptosis and/or necrosis, can be studied in real-time fashion following the exposure of cells to toxic compounds (52).

The major problems with the in vitro systems used currently are their limited, or undefined, capacities for xenobiotic biotransformation. If renal cell lines are to be applied more widely, such as for routine testing purposes, the biotransformation capacities of the cells used should be evaluated and, for specific purposes, cells expressing certain key metabolising enzymes could be constructed using gene technology.

Applications

Risk assessment procedures are applied to provide the best estimate of the expected adverse effects of chemicals in exposed populations. These adverse effects depend on the possible toxic effects of the chemical itself (i.e. the hazard) and the extent of human exposure. Toxic effects of chemicals are influenced by their toxicokinetics and toxicodynamics. Toxicokinetic parameters describe the uptake, distribution, biotransformation, and excretion of the chemical, and thus define the concentration of the chemical at the target site. Toxicodynamic parameters describe the interactions of the chemical with cellular macromolecules, which result in the toxic effects observed. Risk assessment is usually performed (at least in Europe) by collecting all available data on the adverse effects of a chemical in humans, experimental animals, and simpler biological systems, data on toxicokinetics in animals and, if possible, in humans, and information relating to probable exposure levels.

The major potential for in vitro techniques to assist in the hazard identification process for nephrotoxic compounds relates to their use for elucidating mechanisms of chemical-induced acute renal damage, and in identifying certain early endpoints known to be important in chronic processes, such as in carcinogenicity. In combination with other studies on biotransformation, in vitro systems which retain the architecture of the intact kidney may also provide information on selected aspects of toxicokinetics, such as on renal handling and intrarenal bioactivation.

The major limitations to using the currently available in vitro systems for risk assessment purposes are the limited/undefined biotransformation capabilities of these systems and, due to their dissociation from the intact organism and in vivo toxicokinetics, the poor predictivity of these systems for the dose-response curves expected in animals. Moreover, nephrotoxicity due to the production of metabolites in organs other than the kidney, or due to the accumulation of a xenobiotic because of a particular route of administration (for example, the hepatic first pass effect may be important following exposure via the oral route), may not be detected by the exclusive application of in vitro systems for nephrotoxicity testing. Further limitations are the inability of the current in vitro systems to predict complex degenerative processes which result in organ damage. Future developments, including the co-culture of renal cells with non-renal systems, such as hepatocytes or cells of the immune system, may help to overcome some of these problems.

In vitro techniques can, therefore, be introduced into the hazard identification strategy as major tools for elucidating the mechanisms of action of nephrotoxic chemicals. Moreover, in vitro systems can be used to screen series of chemical analogues, to identify those compounds which have the lowest potential to be nephrotoxic, when the toxicokinetics of at least one of the compounds is known or can be estimated by substrate structure-activity relationships.

The specific needs of industry include the availability of metabolically competent systems for nephrotoxicity studies, derived from rodent and non-rodent species, and from humans. These systems should be capable of distinguishing organ-specific and site-specific effects. Co-culture of renal systems with extrarenal metabolising systems (for example, by adding inserts of renal cell cultures to tissue culture plates seeded with hepatocytes) is also needed. For primary screening, precision-cut renal slices, including those from human kidney, could be used to undertake rapid interspecies comparisons of metabolic profiles and chemical-induced toxicity. These slices can be used for up to 24 hours and they enable the investigation of site-specific toxicity. The technique requires less expertise than cell culture, and could now be subjected to prevalidation. Slices, renal fragments, and isolated (in suspension) and cultured renal cells can be used for secondary screening and mechanistic studies. The availability of immortalised cell lines from a range of species would be advantageous. The actual in vitro systems employed will obviously be determined on a case-by-case basis, depending on the specific question being addressed.

Future Developments

It is likely that future technical developments will include the production of:

  1. immortalised cell lines, derived from different renal cell types, which maintain all of the cell-specific functions found in vivo;
  2. heterologous expression systems for xenobiotic metabolising enzymes, which express, in particular, key human enzyme activities; and
  3. more elaborate cell culture systems, such as continuous flow cultures and co-cultures of renal cells with extrarenal cells.

The co-culture of freshly isolated cells and/or permanent cell lines with vascular cells or cells of the immune system could lead to the development of in vitro systems which more closely resemble the situation in vivo. A new method for primary cell culture, incorporating the continuous superfusion of medium, appears to be promising, since it enables the maintenance of primary cultures of different tissues for prolonged periods of time without the loss of cell-specific functions (53). The latter two approaches may be more successful than cell immortalisation in achieving the maintenance of cell-specific functions. The establishment of databases containing in vitro and in vivo nephrotoxicity data suitable for use for comparative purposes is important for the future validation of in vitro systems for nephrotoxicity testing.

Summary and Recommendations

The state-of-the-art with respect to nephrotoxicity testing in vitro can be summarised as follows:

  1. In vitro systems can be used to study the relative toxicities of a series of compounds.
  2. In vitro systems are useful for mechanistic studies.
  3. With optimised phenotypic expression, cell culture systems can be used over a longer period of time than freshly isolated preparations, such as cell suspensions or tissue slices. In theory, this enables compounds to be tested at more realistic concentrations, which may be more relevant for the prediction of subchronic effects in vivo. However, more attention needs to be focused on the refinement of cell culture systems, in order to model toxicity which is manifest in vivo only after chronic administration.
  4. The use of existing renal cell lines is limited by their incomplete characterisation, and by the absence of some renal cell-specific transport processes and biotransformation systems.

The following recommendations are made:

  1. There is a dearth of information in the public domain on correlations between in vivo and in vitro data. A database, drawing on both published and unpublished information from proprietary sources, and preserving confidentiality where necessary, should be established and maintained. Wherever possible, dose-response data from rodent and non-rodent species, and humans, should be included. The need to establish such a database should be addressed by ECVAM.
  2. For screening purposes, the most relevant and sensitive endpoints for the assessment of toxicity should be determined. These will probably vary for different in vitro systems. Using representative nephrotoxins, known to act by different mechanisms, the cascade of toxic events should be elucidated and the early nephrotoxic changes identified. These early changes must be relevant to the final manifestation of toxicity, if they are to serve as indices of nephrotoxicity. Barrier function tests performed using simple and relatively inexpensive electrophysiology equipment may be useful screens with cells grown on microporous supports.
  3. Most of the in vivo nephrotoxicity data available are derived from the rat, but it is difficult to obtain confluent cultures of rat proximal tubular cells. Thus, the optimisation of culture techniques for these cells is required.
  4. The most appropriate non-rodent species for the development of in vitro systems should be determined. Most of the data on primary proximal tubular cell cultures are from the rabbit, a species rarely used for toxicity testing in vivo. Tissue culture systems from the dog, mini-pig, and marmoset should be considered.
  5. The availability of metabolically competent systems is of great importance. The metabolic capabilities of in vitro systems used for studying nephrotoxicity should be evaluated and compared. Strategies should be developed to preserve or restore metabolising enzyme activities in vitro relative to those in vivo. It is also essential that in vitro systems exhibit the same cell-specific transport systems that are present in vivo.
  6. There is considerable literature on the use of primary cultures of proximal tubular cells for pharmacological and toxicological studies. Improved in vitro systems are needed for other cell types, such as distal tubular, glomerular, collecting duct, and medullary interstitial cells. Comparative toxicity studies should be undertaken with primary cell cultures and available cell lines, including those derived from humans.
  7. Co-cultures of renal cells/tissue slices with extra-renal metabolising cells are needed, so that the toxic effects of compounds activated in other organs, particularly in the liver, and then excreted by the kidney, can be investigated. Co-culture systems incorporating cells from the immune system are also required, since some nephrotoxic effects may, either directly or indirectly, involve these cells.
  8. There is a key requirement for in vitro systems which are able to predict chemical-induced nephrotoxic effects which are manifest in vivo only following chronic administration. Thus, efforts should be directed toward the development of stable in vitro systems which are capable of detecting chronic and subchronic effects of chemicals.
  9. Precision-cut cortical and medullary/papillary slices should be evaluated for their usefulness in determining site-specific and species-specific toxicity. The applicability of previously cryopreserved human kidney slices for biotransformation and nephrotoxicity studies should be assessed.
  10. Two alternative strategies for optimising in vitro cell culture systems should be considered: a) stable immortalised cell lines, derived from the principal laboratory test species and from humans, should be developed and characterised with respect to their metabolic competence and transport properties; and b) existing renal cell culture systems should be characterised and manipulated to optimise them relative to the functions and properties expressed by the cells in vivo. The difficulties inherent in the production of suitable immortalised cell lines suggest that progress will be slow and, therefore, both strategies should be pursued concurrently.

References

  1. Anon. (1994). ECVAM News & Views. ATLA 22: 7-11.
  2. Schramek, H., Willinger, C.C., Gstraunthaler, G. & Pfaller, W. (1992). Endothelin-3 modulates glomerular filtration rate in the isolated perfused rat kidney. Renal Physiology & Biochemistry 15: 325-333.
  3. Schurek, H.J. & Kriz, W. (1985). Morphologic evidence for oxygen deficiency in the isolated perfused rat kidney. Laboratory Investigations 53: 145-155.
  4. Bowman, R.H. (1978). Methodology for study of isolated perfused rat kidney in vitro. In Methods in Pharmacology, Volume 4B, Renal Pharmacology (ed. M. Martinez-Maldonado), pp. 385-499. New York: Plenum Press.
  5. Maack, T. (1980). Physiological evaluation of the isolated perfused rat kidney. American Journal of Physiology 238: F71-F78.
  6. Bekersky, I. (1983). Use of the isolated perfused kidney as a tool in drug disposition studies. Drug Metabolism Reviews 14: 931-960.
  7. Guder, W.G. & Ross, B.D. (1984). Enzyme distribution along the nephron. Kidney International 26: 101-111.
  8. Forster, R. (1948). Use of thin kidney slices and isolated renal tubules for direct study of cellular transport kinetics. Science 108: 65-67.
  9. Cross, R.J. & Taggart, J.V. (1950). Renal tubular transport: accumulation of p-aminohippurate by rabbit kidney slices. American Journal of Physiology 161: 181-190.
  10. Berndt, W.O. (1976). Use of the tissue slice technique for evaluation of renal transport processes. Environmental Health Perspectives 14: 78-88.
  11. Smith, J.H. (1988). The use of renal cortical slices from the Fischer rat as an in vitro model to evaluate nephrotoxicity. Fundamental & Applied Toxicology 11: 132-142.
  12. Phelps, J.S., Gandolfi, A.J., Brendel, K. & Dorr, R.T. (1987). Cisplatin nephrotoxicity: in vitro studies with precision-cut rabbit renal cortical slices. Toxicology & Applied Pharmacology 90: 501-512.
  13. Ruegg, C.E., Gandolfi, A.J., Nagle, R.B. & Brendel, K. (1987). Differential patterns of injury to the proximal tubule of renal cortical slices following in vitro exposure to mercuric chloride, potassium dichromate or hypoxic conditions. Toxicology & Applied Pharmacology 90: 261-273.
  14. Ruegg, C.E., Gandolfi, A.J., Brendel, K., Nagle, R.B. & Krumdieck, C.L. (1987). Preparation of positional renal slices for the study of site-specific toxicity. Journal of Pharmacological Methods 17: 111-123.
  15. Fisher, R.L., Hsal, S.J., Sanuik, J.T., Scott, K.S., Gandolfi, A.J. & Brendel, K. (1993). Cold- and cryopreservation of human liver and kidney slices. Cryobiology 30: 250-261.
  16. Vinay, P., Gougoux, A. & Lemieux, G. (1981). Isolation of a pure suspension of rat proximal tubules. American Journal of Physiology 241: F403-F411.
  17. Meezan, E., Brendel, K., Ulreich, J. & Carlson, E.C. (1973). Properties of a pure metabolically active glomerular preparation from rat kidneys. 1. Isolation. Journal of Pharmacology and Experimental Therapeutics 187: 332-341.
  18. McGuinness, S.J. & Ryan, M.P. (1994). Mechanism of cisplatin nephrotoxicity in rat renal proximal tubule suspensions. Toxicology In Vitro 8: 1203-1212.
  19. Ormstadt, K., Orrenius, S. & Jones, D.P. (1980). Preparation and characteristics of isolated kidney cells. Methods in Enzymology 77: 137-146.
  20. Rodilla, V. & Hawksworth, G.M. Isolation and culture of human cortical renal cells with characteristics of proximal tubules. In Methods in Molecular Biology - Human Cell Culture Protocols. New Jersey: Humana Press, in press.
  21. Gordon, E.M., Whiting, P.H., Simpson, J.G. & Hawksworth, G.M. (1987). Isolation and characterisation of rat renal proximal tubular cells. Biochemical Society Transactions 15: 457.
  22. McLaren, J., Whiting, P.H., Simpson, J.G. & Hawksworth, G.M. Isolation and characterisation of human proximal tubular cells derived from kidney cortical segments. Human and Experimental Toxicology, in press.
  23. Heidrich, H-G. & Drew, M.E. (1977). Homogeneous cell populations from rabbit kidney cortex. Journal of Cell Biology 74: 780-788.
  24. Wilson, P.D., Dillingham, M.A., Breckon, R. & Anderson, R.J. (1985). Defined human renal tubular epithelia in culture: growth, characterization, and hormonal response. American Journal of Physiology 248: F436-F443.
  25. van der Biest, I., Nouwen, E.J., van Dromme, S.A. & de Broe, M. (1994). Characterization of pure proximal and heterogeneous distal human tubular cells in culture. Kidney International 45: 85-94.
  26. Chung, S.D., Alavi, N., Livingston, D., Hiller, S. & Taub, M. (1982). Characterization of primary rabbit cultures that express proximal tubule functions in a hormonally defined medium. Journal of Cell Biology 95: 118-126.
  27. Brendel, K. & Meezan, E. (1975). Isolation and properties of a pure preparation of proximal kidney tubules obtained without collagenase treatment. Federation Proceedings 34: 803.
  28. Brendel, K., Hjelle, J.T. & Meezan, E. (1993). Proximal renal tubule preparations isolated from rat or rabbit kidneys without the use of collagenase. Toxicology Methods 3: 87-99.
  29. Genestie, I., Morin, J.P., Vannier, B. & Lorenzon, G. Polarity and transport properties of rabbit kidney proximal tubule cells on collagen-IV coated membranes. American Journal of Physiology, in press.
  30. Zuk, A., Matlin, K.S. & Hay, E.D. (1989). Type I collagen gel induces Madin-Darby canine kidney cells to become fusiform in shape and lose apical-basal polarity. Journal of Cell Biology 108: 903-919.
  31. Schwartzman, M., Ferreri, N.R., Carroll, M.A., Songu-Mize, E. & McGiff, J.C. (1985). Renal cytochrome P450-related arachidonate metabolite inhibits (Na+/K+) ATPase. Nature 314: 620-622.
  32. Wilson, H., Stewart, K.N. & MacLeod, A.M. Glomerular epithelial and mesangial cell culture and characterisation. In Methods in Molecular Biology - Human Cell Culture Protocols. New Jersey: Humana Press, in press.
  33. Chin, T.A. & Templeton, D.M. (1992). Effects of CdCl2 and Cd-metallothionein on cultured mesangial cells. Toxicology & Applied Pharmacology 116: 133-141.
  34. Muirhead, E.E., Germain, G., Leach, B.E., Pitcock, J.A., Stephenson, P., Brooks, B., Brosius, W.L., Daniels, E.G. & Hinman, J.W. (1972). Production of renomedullary prostaglandins by renomedullary interstitial cells. Circulation Research Vol 28 Suppl. II: 161-172.
  35. Tisocki, K. & Hawksworth, G.M. Culture of human renal medullary interstitial cells. In Methods in Molecular Biology - Human Cell Culture Protocols. New Jersey: Humana Press, in press.
  36. Gstraunthaler, G. (1988). Epithelial cells in tissue culture. Renal Physiology & Biochemistry 11: 1-42.
  37. Toutain, H. & Morin, J.P. (1992). Renal proximal tubular cell cultures for studying drug-induced nephrotoxicity and modulation of phenotype expression by medium components. Renal Failure 14: 371-383.
  38. Koehler, G. & Milstein, C. (1975). Continuous cultures of fused cells secreting antibody of predefined specificity. Nature 256: 495-497.
  39. Vandewalle, A., Lehongt, B., Geniteau-Legendre, M., Baudouin, B., Antoine, M., Estrade, S., Chatelet, F., Verroust, P., Cassingena, R. & Ronco, P. (1989). Maintenance of proximal and distal cell functions in SV40-transformed tubular cell lines derived from rabbit kidney cortex. Journal of Cellular Physiology 141: 203-221.
  40. Vandewalle, A., Rafestin-Oblin, M.E., Hagege, J., Geniteau-Legendre, M., Cassingena, R. & Ronco, P.M. (1993). Phenotypic effects of aldosterone and dexamethasone in a SV40-transformed mammalian cortical ascending limb cell line exhibiting mineralocorticoid receptors. Journal of Cellular Physiology 156: 383-394.
  41. Briand, P., Kahn, A. & Vandewalle, A. (1995). Targeted oncogenesis: a powerful method to derive renal cell lines. Kidney International 47: 388-394.
  42. Scott, D.M., MacDonald, C., Brzeski, H. & Kinne, R. (1986). Maintenance and expression of differentiated function of kidney cells following transformation by SV40 early region DNA. Experimental Cell Research 166: 391-398.
  43. MacDonald, C., Watts, P., Stuart, B., Kreuzberg-Duffy, U., Scott, D.M. & Kinne, R.K.H. (1991). Studies on the phenotype and karyotype of immortalised rabbit kidney epithelial cell lines. Experimental Cell Research 195: 458-461.
  44. Lynch, M.J. & Trainer, D.L. (1989). Immortalisation of primary baby rat kidney cells by retroviral-mediated transfer of adenovirus E1A genes. Cancer Research 49: 5429-5434.
  45. Abcouwer, S., Robinson, P.S., Goochee, C.F. & Crow, M.T. (1989). Generation of human embryonic kidney cells with extended in vitro life span through viral oncogene transfection. Bio/Technology 7: 939-946.
  46. Whittaker, J.L., Byrd, P.J., Grand, R.J.A. & Gallimore, P.H. (1984). Isolation and characterisation of four adenovirus type 12-transformed human embryo kidney cell lines. Molecular Cell Biology 4: 110-116.
  47. Perry, S.J., Schofield, M.A., MacFarlane, M., Lock, E.A., King, L.J., Gordon, G.G. & Goldfarb, P. (1993). Isolation and expression of a cDNA coding for rat kidney cytosolic cysteine conjugate β-lyase. Molecular Pharmacology 43 660-665.
  48. Bendirdjian, J.P., Fillastre, J.P. & Foucher, B. (1982). Mitochondrial modifications with the aminoglycosides. In The Aminoglycosides, Microbiology, Clinical Use and Toxicology (ed. A. Whelton & H.C. Neu), pp. 303-324. New York: Marcel Dekker.
  49. Morin, J.P. & Fillastre, J.P. (1982). Aminoglycoside induced lysosomal dysfunctions in the kidney. In The Aminoglycosides, Microbiology, Clinical Use and Toxicology (ed. A. Whelton & H.C. Neu), pp. 303-324. New York: Marcel Dekker.
  50. Godson, C. & Ryan, M.P. (1989). Investigations of gentamicin nephrotoxicity using rat renal brush border membrane vesicles. In Nephrotoxicity: In Vitro to In Vivo, Animals to Man (ed. P.H. Bach & E.A. Lock), pp. 247-252. London: Plenum Press.
  51. van de Water, B., Jaspers, J.J.P.T.G., Maasdam, D.H., Mulder, G.J. & Nagelkerke, J.F. (1994). In vivo and in vitro detachment of proximal tubular cells and F-actin damage: consequences for renal function. American Journal of Physiology 267: F888-F899.
  52. Ryan, M.P., Healy, E., McGlynn, H., Gallagher, A., Monks, R. & McCarthy, L. (1995). Applications of flow cytometry to in vitro nephrotoxicity testing. Cell Physiology & Biochemistry 5: 81.
  53. Minuth, W.M. Dermietzel, R., Kloth, S. & Hennerkes, B. (1992). A new method culturing renal cells under permanent superfusion and producing a luminal-basal medium gradient. Kidney International 41: 215-219.