The Bioartificial kidney
Ogan Abaan, 2001
Background
The kidney, as a part of the human body has vital functions like all the other organs, process blood by removing substances from it and by adding substances to it. The major functions of the kidney can be summarized as follows:
- To regulate the water content, mineral composition, and acidity of the body by excreting each substance in an adequate amount
- To excrete metabolic waste products, urea from proteins, uric acid form nucleic acids, and creatinine from muscle creatinine
- To secrete certain hormones like, erythtopoietin, renin, 1,25-Dihydroxyvitamin D3.
For achieving its functions the kidney is composed of smaller units called nephrons. They have a two-step process: (1) size selective ultrafiltration of the blood by the glomerular capillaries and (2) a reabsorbtion process by the renal tubules using solute specific transporters via active transport mechanisms.
In the case of renal failure, there are two solutions that can be applied. First one is organ transplantation form a donor. However there are many drawback of this strategy like the lack of donors and tissue rejection. Second and more commonly used strategy is the hemodialyasis which first described in 1914 and commercialized in 1956[1]. The kidney was the first organ to be mimicked by and artificial device also the first organ to be successfully transplanted. There is however a problem that the artificial device cannot perform well enough as its biological counterpart.
With the advances in biology in the new millennium, new technologies for improving medical therapeutics are at hand. They are cell therapy, tissue engineering and gene therapy. This review will focus on the available techniques and their drawback as well as the tissue engineered solutions for future applications.
Hemodialysis Vs Hemofiltration
Hemodialysis is diffusion based and clearance is based on molecular size. It mimics the glomerular membrane in the nephron. The clearance of a molecule by diffusion is negatively correlated with the size of the molecule. Hemofiltration on the other hand is convection based and clearance of different sized molecules is similar and dependent on ultrafiltration rate (independent of molecular weight up to 20,000Da), however, replacement fluid is needed. Hemodiafiltration is solute removal by both hemodialysis and hemofiltration simultaneously.[2]
Hemodialysis
Hemodialysis is being used for so many years. During the treatment low molecular waste products arising from the body’s breakdown of ingested protein together with fluid that would be normally passed as urine are removed. Solute clearance profiles were calculated for identical artificial kidney membranes during hemodialysis, hemofiltration and hemodiafiltration. It was shown that the clearance of small solutes depends largely on the dialysate flow rate and is similar when using either hemodialysis or hemodiafiltration. In contrast, clearance of middle molecules, especially low-molecular-weight proteins, depends largely on convective transport induced by high ultrafiltration rates and is maximized when using either hemofiltration or hemodiafiltration. Optimal fluxes for both small solutes and middle molecules can be achieved by using postdilution hemodiafiltration.[3]
The ideal characteristics of a hemodialyzer include high efficiency for solute removal, a flexible capacity of ultrafiltration, minimal or absent proinflammatory effects due to blood-membrane interactions, constancy of performance and safety. This current therapy utilizes synthetic membranes to substitute for the small solute clearance functions of the tubular cells but does not replace the transport, metabolic and endocrinologic functions of the tubular cells. However, a 50-54 years old patient with chronic end stage renal disease on chronic hemodialysis therapy has a five year survival rate of 47% most probably due to the incomplete replacement of the kidney by the hemodialyzer.[4]
The technical process of hemodialysis is subject to countless variations. Apart from variations in blood flow rates, dialysate flow rates, duration of the dialysis episode, and composition of the dialysate, there may also be variations in the type of machinery used and the nature of the dialyzer.[5]
Almost all dialyzers now in use are of the hollow-fiber type. A hollow-fiber dialyzer contains a bundle of approximately 10000 hollow fibers, each with an inner diameter of about 200 micron when wet. The membrane thickness is about 20-45 micron, and the length is 160-250 mm. The walls of the hollow fibers function as the dialysis membrane. Various materials, including cellulose-based materials and synthetic polymers, are used for dialysis membranes.[6]
Collodion, a cellulose-trinitrate derivative, was the first polymer to be used as an artificial membrane and played a central role in further investigations and applications. Cellophane and Cuprophan membranes replaced collodion later, because of their better performance and mechanical stability. However, due to its alleged lack of hemocompatibility, membranes made from unmodified cellulose lost their market share. They have been replaced by modified cellulosic and synthetic dialysis membranes which show a better hemocompatibility than unmodified cellulose membranes.[7] Most of the new membrane materials are also available in high-flux modifications and for this reason suitable as well for more effective therapy modes, such as hemodiafiltration and hemofiltration.
Kerr et al analyzed the performance of six low flux membranes.[8] The dialysis membranes in interest were, cuprammonium (Cu; Terumo), cellulose acetate (CA; Baxter), cellulose diacetate (CD; Baxter), hemophane (HP; Gambro), low flux polysulfone (PS; Fresenius) and polysynthane (PSN; Baxter). All the membranes were in the range of 1.5-1.6 m2 of surface area except PSN that is 1.4 m2. The comparisons were made according to three criteria (1) percent reduction in urea and phosphate (Table 1), (2) instantaneous clearance for urea and phosphate (Table 2), (3) total amounts of urea and phosphate removed in the dialysate per week (Table 3).
Table 1: Percent reductions in urea and phosphate5
Dialyzer |
Urea reduction
ratio % |
Phosphate reduction ratio % |
|
61.0±4.7 |
39.1±12.1 |
CA |
56.9±6.7 |
43.0±8.5 |
CD |
63.0±7.0 |
44.4±10.7 |
HP |
55.9±4.0 |
45.0±6.8 |
PS |
56.6±7.4 |
45.2±7.2 |
PSN |
62.5±5.2 |
40.9±10.3 |
Table 2: Instantaneous clearance data5
|
iCl-urea (ml/min) |
iCl-phosphate (ml/min) |
|
231.3±10.2 |
182.7±12.3 |
CA |
205.1±11.6 |
147.3±9.3 |
CD |
222.0±9.1 |
166.8±15.9 |
HP |
205.5±19.8 |
173.6±14.8 |
PS |
206.4±19.2 |
156.8±23.1 |
PSN |
225.6±18.8 |
164.3±15.2 |
Table 3: Total amounts of urea and phosphate removed in the dialysate per week5
Dialyzer |
Urea per week
(mmol/wk) |
Phosphate per week
(mmol/wk) |
|
2,150±411 |
93.9±32.6 |
CA |
1,707±493 |
88.7±35.7 |
CD |
1,974±575 |
103.0±35.1 |
HP |
2,110±403 |
102.1±35.6 |
PS |
2,105±127 |
85.0±20.3 |
PSN |
2,197±692 |
78.1±10.5 |
From the data obtained, they concluded that there was no single membrane that excelled compared with the other membranes in their in vivo study. The cellulose acetate membrane was perhaps the least performed membrane. Clearances of phosphate during analysis with the different membranes did not necessarily parallel urea clearances suggesting that molecular size and charge selectivity permutations differ considerably according to the nature of the membrane.5 An incident in September 1996 revealed a fact that degradation of the cellulose acetate (CA) material in old dialysis machines was toxic. Degradation products were characterized from retrieved CA dialysis membranes. A series of synthesized CA degradation products was tested in vitro to assess toxicity. Based on the toxicity of the material preparations to the cells, animal tests were performed on selected CA degradation extracts and compared to extracts from actual dialysis membranes.[9]
Small solute removal is obtained by primarily by diffusion. Convection represents an additional mechanism although this is mostly important for larger molecules. The efficiency of a hemodialyzer is therefore dependent on its ability to provide maximal facilitation of diffusion process. Diffusion is affected by blood and dialysate flow rates, temperature, surface area of the dialyzer and thickness of the membrane. Assuming all other factors are constant, the diffusion process is basically dependent on the concentration gradient between blood and dialysate.[10] Ronco et al studied two hemodialyzers with similar overall dimensions but different design of the dialysate compartment. One includes the presence of space yarns between the fibers to prevent the contact of adjacent fibers (figure 1) and the other, a wave design of the hollow fibers (Moiré structure) making adhesion of adjacent fibers less likely to occur.
Figure 1: Schematic representation of the space yarns built in the inter-fiber space to prevent contact of adjacent fibers. The space yarns are also depicted in a scanning electron microscopic section of polysulfone hollow fibers.
Figure 2: Schematic representation of the waved structure of the hollow fibers designed to prevent contact of adjacent fibers (Moiré structure). He external aspect of the hollow fibers is also evident in two scanning electron microscopic picture of a hollow fiber bundle (cellulose triacetate) taken at two different magnifications. In the left bottom panel, the spacing between two adjacent fibers is clearly evident.
Ronco et al concluded that since the fibers are not tightly packed, preferential fluid pathways might be generated. Further, it seems that special configurations designed to prevent a close contact of adjacent fibers may include a significant improvement on the flow distribution. This results in an improved dialysate flow distribution measured by densitometric analysis and an improved urea clearance at similar dialysate and blood flow rates in the case of space yarns or the Moiré design.
There is also the biocompatibility problem of the hemodialysis membranes. Hypoalbuminemia is a marker of high morbidity and mortality in patients with end stage renal disease undergoing hemodialysis. Study by Tayet et al[11] showed a significant increase in serum albumin levels at 3 moths after changing the membrane form a bioincompatible dialysis membrane to a more biocompatible membrane.
Membranes based on cellulose have been long recognized for their suitability in renal replacement therapy; however, over the past decade it has emerged that their biocompatibility is far from optimal. The interaction of blood with the material surface results in the adsorption of proteins to the surface, the adhesion and activation of cells. Since pore size and distribution primarily influences solute and water transport, the purpose of a study by Hoenich[12] et al was to investigate if in a clinical setting it also influences biocompatibility, namely changes in circulating white cells and activation of the complement system since in such a membrane the contact surface with blood is reduced. CuprophanÒ a widely used regenerated membrane produced by the cuprammonium process was used as the baseline. Their study demonstrated that alteration of the membrane morphology by increasing the porosity influences the materials large molecular weight solute transport but fails to modify the materials biocompatibility.
Bioartificial Renal Tubule
A bioartificial tubule uses epithelial progenitor cells cultured on biomatrix coated hollow fiber membranes that are both water and solute permeable, allowing for the differentiated vectorial transport and metabolic and endocrine functions. As the first step toward developing a tissue engineered renal tubule assist device (RAD), Madin Darby canine cells (MDCK), a permanent renal epithelial cell line, were seeded into the lumen of single hollow fibers. The results for the testing of the growth of the cells and functional transport capabilities were successful.[13]
The next step in the development of a renal tubule assist device was to scale up from this single hollow fiber renal tubule to a multifiber bioreactor with renal proximal tubule cells that maintain not only transport properties but also differentiated metabolic and endocrine functions. To accomplish this next step, a reliable nonhuman tissue source for renal progenitor cells is required. For economic and safety concerns pigs can be used as a tissue source. These hollow fiber cartridges with membrane surface areas as large as 0.7 m2 resulting in a device containing up to 2.5 x 109 cells.[14] A surface area decrease from 0.7 m2 to 0.4m2 results in a decrease of the number of cells from 2.5 x 109 cells to 1.4 x 109 cells, In vitro experiments utilizing porcine renal proximal tubule progenitor cells have clearly shown differentiated transport and metabolic functions of the renal tubule assist device unit as summarized in Table 4.
Table 4: In vitro renal tubule assist device
Transport |
Metabolic |
Endocrinologic |
|
Ammoniagenesis-pH sensitive |
1-hydroxylation of vit D3- parathyroid hormone and inorganic phosphate sensitive |
|
Glutathione synthesis- activicin inhibitable |
|
|
|
|
Paraaminohippuric acid- probenecid inhibitable |
|
|
The schematic diagram of bioartificial kidney set-up by Humes et al is shown in figure 3. Specifically, blood is pumped out of a large animal using a peristaltic pump. The blood then enters the fibers of a hemofilter, where ultrafiltrate is formed and delivered into the fibers of the tubule lumens within the renal tubule assist device downstream of the hemofilter. Processed ultrafiltrate exiting the renal tubule assist device is collected and discarded as urine. Heparin is delivered continuously to the blood before entering the renal tubule assist device to diminish clotting within the device. The renal tubule assist device is oriented horizontally and placed into a temperature-controlled environment. The temperature of the cell compartment of the renal tubule assist device must be maintained at 37oC throughout its operation to ensure optimal functionality of the cells. The tubule device unit is able to maintain viability because metabolic substrates and low-molecular weight growth factors are delivered to the tubule cells from the ultrafiltration unit and the blood in the extracapillary space. Furthermore, immunoprotectrion of the cells grown within the hollow fiber is achieved due to the impenetrance of immunoglobulins and immunologically competent cells through the hollow fiber if the encapsulating membrane has a pore size which excludes compounds with a molecular weight greater than 150,000 Da.
Figure 3: Schematic diagram of extracorporeal circuit employed by Humes et al. A synthetic hemofilter followed in series with a tissue-engineered proximal tubule device comprising a monolayer of porcine proximal tubule epithelial cells attached to the lumen surface of hollow fiber membranes.
In the experimental setting of Humes et al dogs were treated either with hemofiltration and the renal tubule assist device or with hemofiltration and a sham control containing no cells. The renal tubule assist devices maintained viability and functionality when connected in series to a hemofiltration cartridge within an extracorporeal perfusion circuit in an acutely uremic animal. Cell loss from the renal tubule assist device during 24-hour perfusion time period showed that less than 105 cells were lost from the device containing more than 109 cells. Treatment with the renal tubule assist device and hemofiltration maintained blood urea nitrogen levels at lower levels than sham controls. In addition, plasma HCO-3, Pi, and K+ levels were more readily maintained near normal values than sham treatment. Table 5 shows the ex vivo properties of the renal tubule assist device.
Table 5: Ex vivo properties of renal tubule assist device
Transport |
Metabolic |
Endocrinologic |
|
Ammonia excretion |
1-hydroxylation of 25-OH vit D3 |
Potassium |
Glutathione reclamation |
|
Bicarbonate |
|
|
Glucose |
|
|
Thus the results clearly show that the combination of a synthetic hemofiltration cartridge and a renal tubule assist device in an extracorporeal circuit successfully replaced filtration, transport, and endocrinologic functions of the kidney in acutely uremic dogs.9 However, one important consideration before the initiation of clinical trials would be the necessity to use human tubule cells rather than porcine cells to eliminate the risk of chronic soluble antigen release from xenogensic tissue into the patient and the potential for immunologic activation.
A potential rate-limiting step in endothelial cell lined hollow fibers of small caliber is thrombotic occlusion, which would limit the functional patency of this filtration unit. In this regard, gene transfer into seeded endothelial cells for constitutive expression and secretion of anticoagulant factors can be envisioned to minimize clot formation in these small caliber hollow fibers.
As it is obvious that there is the need of suitable membranes for the seeding of the cells. Polymers used for the preparation of membranes in conventional bioartificial kidneys such as polyacrlonitrile (PAN) and polysulfone (PSU) have the advantage of a good membrane formation of flat membranes and hollow fibers with controlled pore diameters. The possibility of sterilizing PSU by steam and the ease of surface fictionalization of PAN with various chemical groups might be considered as advantages for a possible application of these polymers in biohybrid organs. However, there is only very limited knowledge about their biocompatibility versus tissue cells. In an experimental setting by Lamprecht et al[15] the growth and morphology of MDCK cells on PAN and PSU asymmetric membranes was investigated. The commercial Millicell-HA cellulose mix-esters symmetric membrane was used as the control membrane.
Both PSU and PAN have immuno-protective properties and would guarantee an immuno-isolation of cells in a biohybrid organ from the host immune system. It was shown by Lamprecht et al that both membranes induced non-toxic action on fibroblasts. They concluded that PAN and PSU membranes provide similar conditions for the attachment and growth of kidney epithelial cells like commercial culture supports. From histology staining it was detected that the highly porous structure of the rough side of the membranes prevented the formation of epithelial monolayers and decreased cell-cell contacts. The transmission electron microscopy demonstrated that MDCK cells attached well to the smooth side of the PAN membranes, but less intimate interaction was observed on the PSU membrane. They explained this phenomenon by he presence of PVP in this membrane, which is known to inhibit cell adhesion. Also formation of close cell contacts and tight junctions was more intensive on PSU membranes and not intensive on PAN.
In another study by Kanai et al the effect of extracellular matrix concentrations of (laminin, fibronectin, and collagen types I and IV) and incubation times (of MDCK cells and KU-2 cells) was evaluated.[16] MDCK line is widely used to investigate the tubular epithelium and KU-2 is a human renal carcinoma cell line. They observed that at each concentration of collagen types I and IV, no difference of MDCK cell attachment onto the matrix among the various times of incubation was observed. At each concentration of laminin and fibronectin, MDCK cell attachment onto the matrix tended to increase as the incubation time increased. Comparison of cell attachments at each incubation time revealed with collagen types I and IV and fibronectin no incubation time dependency of MDCK cell attachment onto the matrix. Only with laminin did MDCK cell attachment tend to decrease as its concentration increased. With all ECM’s, KU-2 cell attachments onto the matrix tended to decrease as the incubation time increased. However, about 24 h after incubation started cell attachments showed the highest absorbance.
Discussion
Up to date there is an extensive research and development on an artificial kidney. Although the first artificial kidney was a membrane at the beginning. The quote was taken form (Kolff’s 1998)[17]. “In 1939, while in Gronighen, The Netherlands, I took a piece of cellophane tubing (50 cm long), put blood in it with 400mg percent urea [metabolic waste in blood plasma], and shook it up and down in a bath with saline. To my surprise, within 5 minutes, nearly all of the urea was removed by dialysis because the surface area was large and both the blood and the dialyzing fluid were in continuous movement. To make an artificial kidney, I could simply multiply this 50-cm-length of cellophane tubing by twenty, and I would have enough cellophane to make clinical dialysis worthwhile. The cellophane tubing was artificial sausage skin.”
After William Kolff's discovery, the first clinically useful artificial kidney was made. It was made of a long cellophane tubing which was wrapped around a steel drum which was spun in a saline bath. Blood was taken from the patient, circulated through the tubing, and then flowed back into the vein. Modifications on this design have been made over time, as there are now cheaper, smaller, more efficient, computer controlled machines. These artificial kidneys are used in hospitals and health care facilities all over the world to help inpatients and outpatients with failing kidneys.[18]
