ORGANIC MOLECULES IN KIDNEY STONES: REVIEW OF SELECTED RECENT PUBLICATIONS
This is a marker paper in which simple urine collection from 10 normal people was followed by centrifugation to remove debris, desalting and concentration, protein separation and very high resolution MS/MS. They found 1543 proteins. Compared to standard proteomic sets for humans urine is high in membrane associated and low in cell component proteins. Specific categories that were very high include protein binding, signal transducers, hydrolases, and calcium ion binding proteins. The entire database is accessible at the lab that did the work.
Calcium oxalate stones
In their introduction these authors note that by 2007 20 proteins had been isolated from human kidney stones, and these were thought to induce crystallization and bind crystals together into large aggregates upon which more crystals might deposit. They had shown that antibodies raised against proteins isolated from stone matrix failed to cross react with urine from normal people but did with urine from stone formers. They report 3 LMW proteins: myeloperoxidase chain A (MPO-A), α defensin, and calgranulin. They did this via analysis of LMW bands from gels of extracted stone matrix using MALDI /TOF techniques.
The stones were from 25 males and 15 female patients. Each had a urine oxalate excretion above 40 mg daily and no other metabolic abnormality; Each stone was washed to remove outer materials, and broken in half. The inner core was cut out from the outer portions, and each was separately pulverized for mineral content. Both portions were predominantly CaOx. Proteins were extracted from the inner and outer cores and used for gels, immunoblots and mass spec analysis (Figure).
The inner core and urine from stone formers contained the three LMW proteins. Osteopontin was found in the inner and outer cores and urine of both normals and stone formers. The LMW proteins were not found in the outer cores or in urine of non stone formers. The authors concluded that because they chose hyperoxaluric patients, not the common hypercalciuric CaOx stone formers the MPO-A, calgranulin and α defensin are response molecules to perhaps oxalate induced renal cell injury. MPO-A is cationic and glycosylated, and known to increase in renal injury states. Defensin is also cationic but nonglycosylated and referred to as anti-inflammatory.
To me the paper merely demonstrates these proteins are present and localized to central stone regions and urine of stone formers. It is most unlikely that their patients had primary hyperoxaluria and therefore that their increased urine oxalate was of such magnitude as to cause renal injury. More likely it is crystals themselves. Another problem is that we do not know if the stones were ever attached to papillae.
This simple paper begins with FTIR documented CaOx stones from unidentified patients, demineralized, and put through conventional ion exchange purification and MW seiving purification with crystal growth inhibition assay to identify promising inhibitory peaks. Maldi-Tof-MS yielded a probable match with human phosphate cytidylyltransferase1, choline, beta which is an enzyme that is rate limiting and regulatory in the CDP choline pathway that leads to phosphatidylcholine (lecithin). This is important in membrane production.
Seven ‘pure’ COM stones from stone registry patients (not detailed in the article) were powdered, proteins were extracted and studied with MS/MS of trypsin digests. They identified 68 proteins of which 50 were considered previously undescribed in stone matrix. Like Mushtag et al they found defensin, MPO and calgranulin, which they categorize as part of the immune response pathway. In addition they found three cell injury proteins and one stress response protein. In addition, they found an array of plasma proteins, albumin being well known, hemoglobin and transferrin, carbonic anhydrase, and a mixture of proteins that include osteopontin. Finally they found cell adhesion, membrane transport, biogenesis, cell signalling and coagulation proteins.
The main contribution here is of quantity. The COM stone contains a large number of peptides, of which quite a few are of inflammatory or immune origin. Some - calgranulins, albumin, clotting proteins, Tamm Horsfall protein, are well known to alter crystal growth and nucleation. Unlike Mushtag et al no attempt was made to relate stone proteins to urine proteins but that may become critical if indeed stones somehow induce the kidneys or urinary tract linings to produce more of certain proteins, such as immune response proteins.
Like others, the authors mention THP. soluble osteopontin, prothrombin fragment 1, bikunin, iTi, calgranulins, fibronectin, and matrix gla proteins as well known inhibitors of crystallization and note that THP, PO and urinary PF1 are in or on stones.
In this study stones from 5 people not otherwise described were used. The crystals of the stones were not mentioned. MS/MS analysis revealed 58 proteins, of which 11 were high abundance and 10 of these present in all three analyses of the extracts from stones. Of these, all but 1 had been previously been identified as binding to CaOx or HA surfaces. Four proteins were prevalent and high abundance on the basis of normalized abundance computation: S100A8 and A9, apolipoprotein A-1, and THP. Using ingenuity analysis, 3 principle pathways were identified: tumorigenesis, immune, and inflammation. The top canonical pathway was acute phase response signaling. The serious weakness of this work is in the stones: unknown crystals. LIkewise for all papers to date the patients are not identified by any of their traits except in one case ‘hyperoxaluria’ of >40 mg/day.
CaOx stones were obtained from 7 men and 3 women not otherwise characterized and their matrix proteins eluted and studied by LC-MS/MS. The stones had been stored at -70oC. As is common selected bands from gels were eluted and studied. They identified 11 proteins. Leukocyte elastase was the most common. Cathepsin G, Azurocidin, and calgranulin; in other words, inflammatory pathway proteins. They did not find osteopontin, perhaps because the bands they selected were too low in MW.
This more recent paper attempts to compare stone matrix materials from patients with different kinds of stones. 15 patients, 14 of them male, provided 17 stones; all had gout or hyperuricemia. Crystals wee identified with X ray diffraction and FTIR. Proteins were extracted and studied with MS/MS. Of interest, osteopontin, protein Z, protein S, and prothrombin were found in COM but not UA stones. Albumin, Hb, uromodulin, the calgranulins, and a histone H4 were present in both CaOx anf UA stones. IgG seemed unique to uric acid stones. Defensin was present in both types of stones. The novelty here is the UA vs the calcium stones.
Whereas most of the crystal active proteins of stone matrix and urine are anions, this paper focuses on cationic proteins from stones. Stones were predominant CaOx by FTIR analysis but numbers of patients and their details are not given. All were surgically removed. CaOx crystal growth inhibition was used as an assay in a standard purification followed by MS/MS identification protocol. Two proteins histone lysine N methyltransferase and inward rectifier K channel, and Wnt-2. were discovered, modelled with relation to the CaOx structure, and mutated for further studies. They found evidence that the three all could inhibit CaOx nucleation and growth and oxalate mediated renal cell injury (presumed from crystals). The article is difficult to interpret exactly, but apparently lysine, phenylalanine, alanine and leucine were found to interact with CaOx crystals.
Thirty stones removed at PERC were used; patients were not characterized except being >25 years old and having no other abnormality but stones. Infection was excluded. Proteins were purified and our crystal growth assay was used. The > 3 kDa fractions were isolated and 2D PAGE performed. 66 spots were found. Of these the 7 most prominent were analysed by Maldi.Toff MS. Disheveled associated activator, glutamate receptor delta - 1 subunit, caspase recruitment domain containing protein, albumin, VP-7 glycoprotein precursor, chymotrypsinogen - A and plasminogen were identified. This paper seems to differ from the others perhaps because of how the protein spots were chosen.
This paper alludes not only to alteration of crystallization by matrix but also to adhesion and bridging to form larger aggregates. But the reference is merely to an old paper by Ryall in 1993 in a review format. For the present research they isolated matrix from 50 stones which were mainly calcium oxalate. Nothing is said about the patients themselves. They made the 50 stones into 5 samples of 10 stones each. Using a growth inhibition assay they found the >3 kDa fraction was active and using a anion exchange made 5 fractions. Of these 1 inhibited growth and nucleation, 2 inhibited nucleation and promoted growth. The first of these was purified by MW seiving into 5 fractions. 3 of the 5 promoted growth; the data look variable to me. Using mass spec they identified ethanolamine phosphate cytidylyltransferase, RAS GTPase activating like protein, and three other odd molecules. Modeling studies confirm that these could have effects on crystals.
Stones were obtained from 9 patients otherwise unidentified. All were predominant CaOx by FTIR. They noted electrophoretic mobilities of certain matrix proteins, such as THP were very heterogeneous and decided to do in solution protease digestion followed by LC/Maldi MS/MS. This differs from most other studies which did peptide digestion and analysis from bands of gels. From the 9 stones done with all peptides, all had some OPN PT THP S100 A8 and A9 and A12 (calgranulins), myeloperoxidase, and albumin. What is striking is the variation of protein abundances among the stones. For the PTF1 fragment abundance varied over 10 fold; for THP variations were slighter. Calgranulins A and B were almost inverse to PTF1 but ran together. Of all the papers this one seems the most interesting. It would appear that selection of the patients, their characteristics and stone proteome will be in the right direction.
This paper concerns urine molecules adsorbed onto COM and COD crystals produced in that urine by inducing increased SS. The good aspects are cleanness of materials; the bad is the lack of biology. CaOx tones do not form as precipitates in urine. The paper is an excellent review article. From it we have that to date of its publication PT fragment 1, albumin, OPN, alpha 2 HS glycoprotein - fetuin, iTI and alpha 1 microglobulin had been shown by laborious techniques to be intra-crystalline components of CaOx crystals. The paper states that OPN is preferentially in COD, iTI is in COM.
For this work they used urine from 4 normals 2 of each sex. Protease inhibitors were not used having been shown by others to be unnecessary for proteomic studies. They made the urine Ca 2 or 8 mM and added oxalate to get crystals. Composition was by FTIR and FESEM. They dissolved the crystals and prepared the adsorbed molecules and did peptide identification with LC-MS/MS. They also made 2 D gels of different batches of crystal matrix and analysed peptides from spots that seemed dominant. From COM they found OPN, PF1 S100 A9 Ig S100 A8 Hornerin Histone H4 Protein Z and a few others. From COD they found OPN, the Ig S100 A9 annexin, serine protease inhibitor, kininogen, iAI and a large mix of other proteins. The proteins common to both are in this table. Of interest they did not find THP in any of the batches nor PTF1.
Perhaps more than most this paper exposes the complexity and uncertainty of the stone crystal matrix problem. Stones themselves may alter the abundances of the urine proteome and therefore of stone matrix. Proteins inside crystals, for that is what is asserted here, would have altered growth and nucleation rates and could favor one form of CaOx over another as noted by Nancollis.
Calcium phosphate stones
As opposed to the CaOx papers which catalog matrix proteins, this one studies phosphate containing stones whose carbon atoms - apart from carbonates - are all organic and phosphorus is all in a mineral phase. Because P31 and C13 are stable and have favorable NMR properties NMR can be used to identify P31 and C13 atoms in close proximity, meaning organic molecules close to or within mineral. The close order of the proximity, 0.5 nm when present would suggest an intimacy mediated by intermolecular forces such as ion pairing and hydrogen bonding.
Using the technique, 10 stones - not identified as to patients - showed close proximity of organic and crystal, especially marked for glycosaminoglycan (GAG) acidic groups. By comparison, the incorporation was far more marked among HA vs. struvite CaP stones. To the investigators this implies that the phosphate crystals and organic molecules are biopolymer composites not simple mixtures of discrete organic materials surrounded by mineral. In stones where the carbon is from uric acid or calcium oxalates, there was no evidence for close intermolecular forces, suggesting these phases are distinct from one another.
The work is important in that HA of plaque and of the nascent overgrowth on plaque may well form in close relationship to GAGs or collagens, not merely close but as biopolymers. It shows a technique that is best applied to, as an example, human biopsy materials of plaque with stone overgrowth. It is the only work so far that suggests that GAGs or other kinds of proteins might bind crystals together, incidentally.
HA stones are composed of a ‘seemingly amorphous HAP phase..’ made of layers and blocks with abundant organic material. Spherules are disseminated throughout the stones; they are 10 - 20 u and layered like onions with concentric white and dark layers. The fine structure, by SEM, appear to be made of finer spherulites which are thought to form by sedimentation in a relatively stagnant part of the kidney or urinary tract. This work uses AFM to analyse the ultrafine structure. A single stone was studied. They show a layered section with alternating thick light and thin dark lamina. By AFM the dark layers are elevated above the white regions, and are composed of columnar crystals. The white layers are agglomerates. Each agglomerate is made of hundreds of smaller spherulites of about 10 nm, and organic material is abundant.
Their view is that the dark and white layers are both made of spherulites agglomerated into spherules, with columnar crystals predominating in the elevated sections. Because of the roundness of the microspherulites and their larger bundles that make up spherules, they conclude formation must provide material equally from all sides and therefore they were formed floating in the urine. In vitro, HAP solutions produce ‘Posner’s’ clusters of 15 ions - 9 Ca, 6 PO4, about 0.7 nm each. These are very transient and rapidly coalesce to form larger aggregates. Brownian motion serves to promote their joining. As clusters grow their motion falls so they join more slowly which leads to stability around the 10 nm range. However they can aggregate and be retained in ‘cavities with poor urodynamics’ to form a stone. The agglomerates are covered with organic material which ‘segregates’ them from one another, taken as evidence for settling or sedimentation in urine.
A quantitative approach is taken using Stoke’s law for settling, but nephrons do not function as if gravity driven, so I do not accept the formulation. Flow is always occurring in kidneys. But their ladder of size is of value: nanoclusters (Posner’s) 0.7 nm; spherulites, 7-10 nm; agglomerates, 150 - 300 nm; bulk phase 10-20 um. THey estimate 11 minutes is enough to form 500 10 nm spherulites. An IMCD lumen is perhaps 20 um, a BD much larger.
This paper appears to neglect the organic material altogether as important in stone structure and place all emphasis on crystal aggregation. It also assumes some kind of poorly stirred places in kidneys and urinary tract. It is hard to place in perspective in that urine matrix molecules adsorb onto HAP and alter its growth. Likewise, the massive tubule plugging of HA stone formers is presumably related to their stones.
CaOx vs. CaP stones
This work takes off from our own. We have proposed that CaOx stones grow outside the kidneys over plaque whereas CaP stones involve tubule HA plugs with obvious inflammation and cell injury. A prediction they make from our findings is that CaP stones should include a higher proportion of inflammatory proteins than do CaOx stones and they tested that using proteomic analysis of 29 CaOx and 18 CaP stones obtained from their stone bank. Of these, 10 COM, 4 COD, 9 HA and 3 BR met criteria for crystal analysis. MS/MS was used to identify proteins as in other papers of this type.
No differences in inflammatory proteins were found. The COD stones were different from COM and Br were quite different from HA but not in regard to inflammation. He has our work wrong in that plaque seems to be occurring within tubules, incidentally. What shows up most clearly is the presence of the response proteins. Calgranulins A and B THP antitrypsin, defensin, osteopontin, plasminogen, all are of high abundance. One thinks that perhaps subgroup analysis might help as it was not done.
This is a theoretical working out of formation of CaOx and HA in solution by an expert with excellent practical sense. Of particular interest he notes that amorphous CAP (ACP) is a hydrated tricalcium phosphate forming nano-sized assemblies. These associate into larger assemblies until they transform to mature crystallites of HA. These are of course the nanoparticles in the Grasses paper, but they did not mature. Perhaps the matrix molecules prevent their maturation. He shows an experiment in which brushite is added to a solution supersaturated with COM; the brushite dissolves leading to nucleation of COM on whose surface the brushite seeds attach, or adhere. He points out that in stones this would lead to would be COM with brushite cores. He points out that matrix molecules can hinder the formation of a more thermodynamically favored phase so that crystallization might be altered.
Calcium stone formers
Urine proteins apart from albumin are generally low molecular weight with the exception of THP. They divided what appear to be idiopathic calcium stone formers into those with high (>4.3) or low fasting non albumin urine protein (in mg). Among patients with stones in their kidneys traits of those higher vs. lower included higher systolic blood pressure, urine uric acid, citrate, and pH. Volume was higher. Among those with vs. without stones in their kidneys, differences of the protein were not significant. Likewise for those with higher vs lower urine pH. The protein excretion seemed to rise with urine volume. Although the proteins of matrix probably reside within the fraction he is studying, it is difficult to place this paper into the general theme of stone matrix function.