Luận án tiến sĩ: Kinetics studies of reactions at solid - liquid interface: Simulation of biomineralization

Symbol a i ion activity C ds average counts of adsorption/desorption of N2 gas for the sample C dN2 desorption counts for the known volume of N2 gas during calibration C eff C-effective:

Theory of Crystal Nucleation & Growth 1

Nucleation 2

The formation of a new phase in the body of the ambient phase, such as gas or liquid, is one of the most fundamental aspects of phase transitions, in particular, crystallization The potential barrier which a system must overcome in order to create a (crystalline) nucleus in the ambient phase, and which determines the rate of nucleation, is defined by the interfacial energy If under a certain condition the probability of creating a nucleus is uniform throughout the system, nucleation is defined as homogeneous nucleation Otherwise, it is defined as heterogeneous nucleation 1 Heterogeneous nucleation normally occurs on solid or liquid surfaces, microclusters, dusts, macromolecules or other foreign bodies In most system these foreign bodies are detected inevitably Therefore, in most cases, nucleation is heterogeneous rather than homogeneous 2 During heterogeneous nucleation, the properties of these foreign bodies represent an additional factor upon which the nucleation barrier and rate depend In this section, heterogeneous nucleation will be discussed; homogeneous nucleation can be treated as an upper limit of heterogeneous nucleation

The nucleation of crystals from aqueous solution is highly dependent on the supersaturation with respect to the crystal phase The supersaturation in aqueous solution of sparingly soluble electrolytes, S, can be written in terms of the ionic product (IP) of the crystal lattice ions and the solubility product constant (Ksp) 3

In eq 1-1, ν is the number of ions contained in the formula unit of the precipitating phase ion activity A typical solubility isotherm for AαBβ is illustrated in Figure 1-1 It can be seen that the supersaturation region may be divided into three quite distinct parts, S1 When S1) solution may be either metastable (1> k d , then K G ≈ k d and growth is controlled by diffusion process When k d >> k o (with K G ≈ k o ), growth is controlled by its orientation process

Based on thermodynamics, Volmer 25 proposed a crystal growth mechanism based on the existence of an adsorbed layer of solute atoms or molecules on a crystal face When growth units arrive at the solid-liquid interface, they would not be immediately fully incorporated into the crystal lattice, but would migrate over the face of the crystal by forming a layer of adsorbed units

Ions and molecules from solution will adsorb on the lattice at the “active sites” of the crystal, where the attractive forces are greatest, until the whole plane face is completed The step-wise addition of growth units across a crystal results in a flat layer devoid of active sites; therefore, a two-dimensional nucleus must be produced on the adsorption layer to continue crystal growth (Figure 1-7)

From a non-thermodynamic point of view, Kossel 26 proposed that a growth unit must be correctly oriented on the crystal surface before it could be incorporated into the lattice The Kossel model of a growing crystal face can be illustrated as Figure 1-8 The surface consists of flat regions called terraces and raised partial layers called steps The steps themselves are also incomplete, containing kinks and vacancies Nuclei are most readily adsorbed onto the crystal at kinks and the kink will then move along the step edge until the face is complete A new face is then initiated by a surface nucleation event and the process is repeated A crystal should grow fastest when its faces have a large number of kinks because molecules that attach there make more bonds to neighboring molecules than the ones that attach to the terraces or to flat step edges

In the mononucleation growth model, growth begins by the formation of a two- dimensional nucleus on a flat crystal surface, which then spreads laterally across the crystal face at an infinite velocity (Figure 1-9a) Unlike the mononuclear model, the nucleus spread speed is zero in the polynucleation model The growth occurs through the formation of a two-dimensional nucleus on a flat crystal surface, until the entire layer is covered (Figure 1-9b)

The “birth-and-spread” (BS) model allows the formation of two-dimensional nuclei and their spreading across the surface at a finite rate (Figure 1-9c)

Burton-Cabrera-Frank (BCF) Theory

In 1951 Burton, Cabrera and Frank 29-31 developed a kinetic theory of growth in

Figure 1-7 1 A mode of crystal growth without dislocation: (a) migration towards desired location; (b) completed layer; (c) surface nucleation

Figure 1-8 1 A schematic of the Kossel model of crystal growth K refers to kink sites; S represents steps; T a terrace; E a pit; G loosely adsorbed growth sites; and N a nucleated crystal.

Figure 1-9 27 Three types of current two-dimensional crystal growth mechanisms: a) mononucleation; b) polynucleation; and c) birth and spread which the crystals were not ideal, three-dimensional, flat, cubic blocks, but instead contained numerous imperfections or dislocations The adsorption of a growth unit to the crystal lattice at the dislocation would result in the production of continuously self- perpetuating new step and kink sites (Figure 1-10) They were able to calculate the growth rate for any supersaturation by applying Boltzmann statistics and predicted kink density by assuming that surface diffusion is an essential step in the process:

R= (1-23) where R is the rate of crystal growth and A and B are complex temperature dependent constants The crystal growth rate is highly dependent on σ, at low supersaturations, equation 1-23 can be simplified to the parabolic relationship Rα σ 2 , while at high supersaturations a linear dependence would be expected (Figure 1-11)

Most likely, however, crystal growth is the result of a combination of processes some of which are listed above In addition, changes in temperature, particle size, and fluid dynamics will influence both crystal growth and dissolution rates

1.2.2 Measurement of Crystal Growth Rate

The crystallization rates of sparingly soluble salts may be expressed in several ways, the most common of which is to write the rate as a function of driving force Nielsen 32 suggested an empirical rate law n k n

Where k n is the rate constant and n is the reaction order Plotting logR verses log σ will yield the value for n which can be useful in interpreting the controlling mechanism as shown in Table 1-1

Figure 1-10 1 Schematic of Burton-Cabrera-Frank model of spiral growth

Figure 1-11 1 Plot depicting the dependence of spiral growth rate on supersaturation R is the rate of crystal growth and σ the supersaturation

Table 1-1 Effective order of reaction, n, and suggested mechanism n Proposed mechanism

2 Surface diffusion, integration, or a combination of both

Experimental Methods and Materials 34

General Methods and Materials 35

All solutions were prepared using reagent grade chemicals and triply distilled deionized water (TDW) and were filtered twice using 0.22 àm Millipore filters (Millipore đ , Bedford, MA) before use Calcium chloride (CaCl2) solutions were standardized by titrating against disodium ethylenediamine-tetra acetate, EDTA, with an NH4 +/NH3 buffer and Erichrome Black-T indicator 1 Potassium dihydrogen phosphate (KH2PO4) was prepared fresh every week to avoid bacteria growth and was prepared by weight from 110 o C dried solids Potassium hydroxide (KOH) was prepaired in a nitrogen atmosphere using carbon dioxide free TDW Hydrochloric acid (HCl) solutions were prepared from 0.1 molL -1 DILUT-IT solutions and were titrated against a standardized KOH solution with methyl red indicator Potassium oxalate solutions (K2C2O4) were passed through an anion exchange resin (Dowex ® 1-X8 50-100 Mesh) in the hydroxide form and the eluted base was titrated with hydrochloric acid solution using methyl red as an indicator

Three buffer solutions (Table 2-1) were used to calibrate the glass electrode They were prepared according to National Institute Standards and Technologies 2, 3

Aliquots of reaction solution were taken throught each experiment and analyzed for calcium and phosphate Calcium concentrations were determined by atomic absorption spectrometry at λ= 422.2 nm utilizing a Perkin Elmer 3100 spectrometer with an air/acet-

Table 2-1 Buffer solutions used for calibrating pH glass electrode pH

4.008 4.028 5.0E-2 mol kg -1 potassium hydrogen phthalate 6.865 6.843 2.5E-2 mol kg -1 Na2HPO4 + 2.5E-2 mol kg -1 KH2PO4

7.413 7.386 3.0E-2 mol kg -1 Na2HPO4 + 8.7E-3 mol kg -1 KH2PO4 ylene flame source UV-Vis spectrometry was used to determine the phosphate concentration as a phosphovanadomolybdate complex at λ= 420.0 nm The vanadomolybdate developing solution consisted of three components, A, B, and C, which were mixed in equal volumes immediatedly before use

50.0g lanthanum oxide (La2O3) + 200.0 mL concentrated HCl diluted to 1 L with TDW Experimentally, it was determined that solution A should be made with at least 75% concentrated HCl to ensure proper phosphate development and a more translucent phosphate blank

50.0 g ammonium molybdate [(NH4)6Mo7O24 4H2O] diluted to 1L with TDW This light sensitive solution was stored in the dark

2.0 g ammonium vanadate (NH4VO3) + 200.0 mL concentrated nitric acid (HNO3), diluted to 1L with TDW

Standard solutions were prepared with stoichiometric Ca/P molar ratios When preparing standard and sample solutions for analysis, 1/5 of the total solution volume consisted of the developing solution In addition, a blank solution was prepared with a matching electrolyte and developing solution background Samples were allowed to develop for 2-212hours to ensure optimum phosphovanadomolybdate complexation and chemical analysis was performed within the hour to prevent solution degradation

To determine the free lattice ion activities, it is necessary to take into account the formation of ion-pairs and complexes in the solutions This was achieved by using expressions for mass balance and electroneutrality with the appropriate equilibrium constants, with successive approximations for the solution ionic strength as described by Nancollas 4 The mean activity coefficient, γ ± , was calculated using the extended form of the Debye-Hückel equation, proposed by Davies: 5

= + − γ± (2-1) where A is a temperature dependent constant, z + and z - are the charge of the cations and anions respectively and I is the ionic strength, defined by

I (2-2) c i is the concentration (mol L -1 ) of the ith ionic species

A speciation program, developed by Arman Ebrahimpour in this laboratory, was used to calculate the driving force, σ Important equilibrium constants, used in program are listed in Table 2-2

DCPD crystals were prepared by the slow addition at 30 o C of 250 ml of 0.02 M disodium hydrogen phosphate to a mixture of 125 ml of 0.40 M calcium chloride and 250 ml of 0.18 M potassium dihydrogen phosphate Nucleation commenced at pH 5.5 which was maintained approximately constant during precipitation by controlling the addition of

Table 2-2 Calcium Oxalate and Calcium phosphate ionic equilibrium equations and corresponding constants at 37 o C

Reaction Equilibrium Constant at 37 o C Reference

Ca 2+ + H2PO4 - ' CaH2PO4 + 27.90 L mol -1 9

TDW The crystals were then dried in vacuo at room temperature X-ray diffraction spectra agreed well with the published data 10-12 The DCPD seed crystals had a calcium to phosphate molar ratio of 1.00 ± 0.04 and a specific surface area (SSA) of 3.92 ± 0.5 m 2 g -

1 as measured by a single point BET nitrogen adsorption (30/70 He/N2, Quantasorb Sorption System, Quantachrome Corporation)

HAP seed was prepared by an hour long drop-wise addition of two separate solutions 1) 600 mL of 0.5 mol L -1 calcium nitrate tetrahydrate [Ca(NO3)2.3H2O] and 2)

600 mL of 0.3 mol L -1 ammonium phosphate dibasic [(NH4)2HPO4] to 500 ml of a 10% ammonium hydroxide (NH4OH) solution at a temperature slightly greater than 90 o C The reaction was taken at constant pH>10 by adding NH4OH and the solution was allowed to reflux for approximately 5 hours The precipitate was then filtered and washed with TDW and stored as slurry in TDW for more than a year before use The HAP seed crystals had a calcium to phosphate molar ratio of 1.68 ± 0.01 and a specific surface area (SSA) of 24.1± 0.5 m 2 g -1

2.1.5 Calibration of pH and pCa electrodes

An Orion 720A pH meter was used to measure ion activities during the experiments Both the glass electrode (Orion 91-01) and the calcium ion selective electrode (Orion 93-20), were coupled with an Ag/AgCl double junction reference electrode (Orion 900100) filled with a 4M KCl solution saturated with AgCl H + activity was measured by a glass electrode The calcium ion activity was measured using an Orion calcium ion selective electrode which was calibrated by adding at least three aliquots of calcium chloride stock electrode 13 The measured electrode potentials were given by the Nernst Eq (2-3)

E 0 ( )ln (2-3) where E is the measured electrode potential, E 0 is the difference in standard potentials between the glass/calcium ion selective electrode and the Ag/AgCl reference electrodes,

R is the gas constant, T is the temperature, F500 C/mol, is Faraday’s constant, and

M z a is the ion activity A calibration curve was constructed of the measured potential as a function of the ln of the ion activity

The specific surface area (SSA) of the each seed crystal was determined by a single- point BET (Brunauer, Emmett and Teller) method using a Quantasorb continuous flow instruement (Quantachrome Corporation) A 20/80 Nitrogen/Helium mixture was passed through a U-tube cell containing the sample for 30 minutes, cooled by liquid N2 in order to trap any impurities Adsorption of nitrogen gas to the sample was measured by thermal conductivity on submerging the cell into liquid nitrogen Conversely, desorption of N2 gas was measured upon removing the cell from the liquid N2 The instrument was calibrated using a known volume of N2 gas and the SSA (m 2 g -1 ) was calculated by using

= 2 (2-4) where K SSA is a constant dependent on temperature and atmospheric pressure, V is the N 2 volume of N2 used to calibrate the apparatus, m s is the sample mass, and C ds and dN 2

C are the average of counts of N2 desorption/adsorption for the sample and desorption calibration counts of N2, respectively

The solid phases, separated by filtration (0.22 àm Millipore đ filters), under vacuum, were sputter-coated with a thin carbon deposit to provide conductivity, and then examined using a field-emission SEM (Hitachi S-4000 FESEM), typically at 20 or 30 KeV

X-ray patterns of the solid phases were recorded using a Siemens D500 diffractometer with Cu Kα radiation The diffractometer was equipped with a diffracted beam monochromator and a scintillation counter The diffractometer was operated in a step scan mode with a step size of 0.02° 2θ, and the count time at each step was 1 s The diffraction effects were recorded in the range between 2° and 80° All images were taken by Anja Dosen of the Geology Department, University at Buffalo

In situ AFM The images were collected in contact mode using Multimode AFM with Nano III controller (Vecco) Images were acquired in height and deflection modes using the lowest tip force possible to reduce the tip–surface interaction The brushite seed crystal was anchored inside the fluid cell and supersaturated solutions passed through it while the images were taken All images were collected by Molly Darragh and Dr Jennifer L Bioconidi, of Lawrence Livermore National Laboratory.

Kinetics Measurements 42

The constant composition method has been used extensively to study the kinetics of crystallization and dissolution of a variety of sparingly soluble salts The degree of supersaturation is maintained constant during the reaction allowing for the formation of relatively large amounts of newly precipitated solid phase for physical-chemical characterization The titrant solutions, added in order to compensate for removal of lattice ions during crystallization reactions, are prepared with appropriate stoichiometry matching that of the precipitating phase and allowing for dilution due to addition from multiple burets By using very dilute titrants, it is possible to determine the rates of reaction even when the driving force is very low – conditions impossible to study using converntional free drift methods in which the lattice ion concentrations decrease with time By using the CC method, it is also possible to select particular points on the solubility isotherm diagram to limit the number of precursor phases that can form, thereby making it possible to study the influence of additives on reactions with an increased precision 15-17

A schematic of typical CC instrumentation is given in Figure 2-1 Experiments were performed in polyethylene covered double-walled Pyrex vessels with magnetic stirring (450 rpm), thermostatted by a circulating water bath at 37.0 ± 0.1 o C Pre-saturated nitrogen was passed through the reaction solutions to exclude carbon dioxide The supersaturated reaction solutions were prepared by slowly mixing calcium chloride and potassium dihydrogen phosphate with sodium chloride to maintain the physiological ionic strength, I = 0.15 mol L -1 Super- or under-saturation solution was achieved by adjustment of the pH with either KOH or HCl, respectively

The growth reactions were initiated by the introduction of brushite seed crystallites, no seed was used for the CC nucleation experiments The electrode potential was constantly compared with a preset value, and the difference, or error signal, activated two motor-driven titrant burettes to maintain a constant thermodynamic driving force

Throughout an experiment, aliquots of reaction solution were periodically removed and filtered using 0.22 àm Millipore filters The calcium and phosphate concentrations were analyzed by atomic adsorption and UV-Vis, respectively Constant composition was confirmed to within 2% for both calcium and phosphate

The following general formula was used to calculate the titrant concentrations: x x x H PO OH

If W 1 , W 2 , W 3 , and W 4 denote the total concentrations of CaCl2, KH2PO4, KOH and NaCl in the working solution, respectively and T 1 , T 2 , T 3 , and T 4 represent the titrant concentrations of CaCl2, KH2PO4, KOH and NaCl, respectively Total calcium is given by Eq (2-5) t b T t

1 (2-5) where V T is the total volume of the working solution, V t is the volume of titrant added (ml), dn is the number of moles precipitated, n b is the total number of burets

The titrant concentration, C eff is defined as the number of moles of deposited crystal per liter of added titrant and may expressed as eff dnV t

T = b + b − (2-11) Substitution of Eq.(2-7) into Eq (2-11) gives Eq (2-12)

T = b + b + b − − (2-14) From Eq (2-9), (2-12), and (2-14), Eq (2-15) is obtained

In summary, if x = 0 in Eq (2-7) for HAP; if x = 2, for OCP

Similarly, for CaC 2 O 4 H 2 O growth experiment, the concentration of Ca 2+ and

C 2 O 4 2- ion after titrant addition is given by Eq (2-16): t i T i i V n V dn V T V

Eq (2-18) is also given to maintain the ionic strength eff NaCl

T =2 −2 (2-18) Table 2-3 gives the titrant equations for HAP, OCP, COM growth, and for DCPD dissolution

During constant composition experiments, the volume of titrant added is recorded as a function of time The rate of crystal growth, (mol m -2 min -1 ) may be expressed as: seed t eff m SSA dt C dV

(2-19) where dV t dt is the rate of titrant addition (L min -1 ), m s is the seed mass (mg), and SSA is the effective surface area (m 2 )

Changes in titrant volume as a function of time provide important information on the surface area available and the extent of crystal growth Assuming simple, uniform three- dimensional growth, the seed specific surfaces areas at any time, SSA t , is expressed by

SSA = (2-20) where SSA ini is the initial seed specific surfaces areas, m t is the total mass of crystal at any time, then m t =m seed +V t ⋅C eff ⋅MW (2-21)

MW is the molecular weight (g mol -1 ) of precipitating phase By substituting Eq (2-20) and (2-21) into Eq (2-19), the corrected crystal growth rate R g , is obtained

Table 2-3 Titrant equations for Hydroxyapatite, Dicalcium phosphate dihydrate,

Octacalcium phosphate, and Calcium oxalate monohydrate growth and dissolution experiments

HAP [Ca 5 (PO 4 ) 3 OH] eff KOH b KOH eff P b P eff NaCl b NaCl eff CaCl b CaCl

+ eff KOH b HCL b KOH eff P b P eff KOH b NaCl b NaCl eff CaCl b CaCl

2H 2 O] KOH KOH eff eff P P eff NaCl NaCl eff CaCl CaCl

= eff KCl KOH eff P P eff NaCl NaCl eff CaCl CaCl

KH eff NaCl NaCl eff CaCl

H 2 O] Ox Ox eff eff NaCl

In biological mineral systems, reaction solutions could be supersaturated/ undersaturated with respect to more than one phase that promotes either simultaneous growth or dissolution of other phases Investigating the simultaneous kinetic processes of these mixed phase systems was difficult and has severely limited previous studies of the formation kinetics of biological minerals The dual constant composition method can model these multiple kinetic processes selectively and determine an accurate reaction rate for each phase

The simultaneous growth or dissolution of BA and BC crystals with a common ion was monitored through the use of two ion-selective electrodes reversible with respect to

B and C and controlling BA and BC titrants, respectively The dissolution of phases BA and BC resulted in a change in each ion concentration and registered as a deviation in the electrode potential difference, ∆E, from a set point ([X]s to [X]):

In Eq (2-23), z was the charge on the ion to which the electrode was reversible and F, R, and T were the Faraday constant, the gas constant and the temperature (K), respectively Titrant addition was triggered by the impulsomat when a potential difference exceeded the response threshold, ∆U, to compensate for the increasing ion concentration resulting from the dissolution of BA and BC The magnitude of the relative concentration before the response threshold was exceeded and the titrant added was defined as

A value directly proportional to the valence of the specific ion being measured by the electrode

The time interval needed to trigger the BA titrant addition was given by the equation:

1 ε (2-25) where V was the volume of the reaction solution and R BA and R BC were the dissolution reaction rated (mol.sec -1 ) of the mineral phases BA and BC respectively The BA titrant was then added to re-establish [B] to the former setpoint value, [B]s, and was determined by the equation:

Where C eff,BA was the effective concentration of the BA titrant Consequently, the concentration of B was maintained constant with the addition of the BA titrant However, the BA continued to dissolve After another time interval, dt 2 , the change in the [C] was sufficient to initiate the addition of BC titrant as defined by the equation:

= ε (2-27) and the volume of titrant BC was given by the equation:

Where C eff,BA was the effective concentration of the BC titrant The BC titrant addition was controlled directly by the electrode sensitive to the change in [C] as if BC was the only phase dissolving in the system As a result, the dissolution rate for the BC phase

The signal magnitude for the addition of the BA titrant in the time interval of dt2 was equivalent to

C ε ε but varied since the signal was dependent upon the electrode sensitivity with respect to the B mineral component However, the introduction of the BC titrant into the reaction solution also affected the BA titrant addition by reducing the amount of BA titrant required Therefore, the total amount of BA titrant needed in the time interval, dt 2 , was given by the equation:

BA C dV C dt dV dt dV

By rearranging Eq.2-30 and substituting Eqs 2-25 and 2-26 the rate of dissolution for the

BA phase was determined from the equation:

(2-31) where dV BA /dt 2 is slope of the BA titrant-time plot

After initiation of the reaction by seeding with BA and BC crystals, titrant BC would not be added until [C] decreases by a factor of ε which is needed to trigger the addition of the BC titrant This addition would cause a pulsed increase in [C] back to the set value This cycle would then be repeated throughout the reaction Thus the variation of [C] in the working solution would be similar to that of a single CC experiment, and would be controlled to within 0 and –ε The concentration of the common ion, [B], on the other hand, would fluctuate between –ε and +ε The upper boundary will be reached only likewise would be within –ε and +ε (Figure 2-2b) Similarly, for the simultaneous tgrowth of BC and dissolution of BA it can be shown (Figure 2-2b) that while [C] is maintained within 0 and –ε, [A] and [C] can vary between 0 and 2ε

Figure 2-2 Relative concentration variations during simultaneous (a) growth of BA and

BC and (b) dissolution of BA and growth of BC The concentrations of A, B, and C are represented by dash, dot and solid lines, respectively.

The Potential Calcification of Octacalcium Phosphate on Intraocular Lens Surfaces 56

Background 57

What is an intraocular lens?

An intraocular lens (IOL) is an artificial lens inserted during cataract surgery to replace the eye’s natural lens An intraocular lens also may be surgically inserted as a method of vision correction that leaves the eye's natural (crystalline) lens intact An IOL consists of 2 parts, a main part that consists of the optic to restore the best vision possible and a peripheral part (haptic), see Figure 3-1

The ideal intraocular implant should be chemically inert It must be stable and not modified when it is in contact with living tissue It also should be biocompatible, non- carcinogenic, non-allergenic, machinable, transparent, and implantable by a small surgical incision It must ensure blockage of UV radiation and thermally stable, i.e autoclavable (for sterilization) 1

In 1949 the ophthalmologist Harold Ridley examined a pilot who had just received in his cornea a splinter from the windshield of his cockpit, made of PMMA (polymethyl methacrylate) Although the accident had taken place several days ago, Dr Ridley found that the PMMA has been tolerated very well by the tissue Taking into account this medical data, H Ridley manufactured the first intraocular lens 2-8 The second generation of lenses (1952 – 1962) were rigid, closed loop lenses, supported by the anterior chamber angle The third generation of IOLs (1953 – 1973) was independently produced by Epstein and Binkhorst It used the papillary part of the iris diaphragm for anatomical fixation For an anterior chamber lens to be safe and effective there should be minimal contact with the drainage angle, stability within the anterior chamber with no movement

Figure3-1 Examples of IOL (left, OII Aqua-Sense TM ; right, OII Aqua-Sense III TM ) in the angle, no iris chafing and no endothelial touch Modern anterior chamber IOLs (1970 to present day) were designed to achieve this by using flexible open loop lenses made of PMMA The first posterior chamber IOLs (1975 to present day) were made of PMMA with either PMMA, polypropylene or polyamide haptics

Pathological explanation about formation of cataract

The crystalline lens of the eye is positioned behind the iris with its posterior aspect embedded in the vitreous body (Figure 3-2) 9 The lens plays a passive role in the process of accommodation by which light rays, which have passed through the cornea and aqueous humor, are focused upon the retina The molecular makeup of the lens is also unique in that it is two-thirds water with one-third protein; other constituents represent only about 1% of the total lens wet weight

The lens comprises two types of cells: the single layer of cuboidal epithelial cells on its anterior surface and the elongated fibre cells which make up the bulk of the tissue (Figure 3-3) Epithelial cells proliferate at the anterior surface and migrate laterally to the equatorial elongation zone where they differentiate into fibre cells The newly formed fibre cells are laid down over existing fibre cells, which are compressed into the center of the lens, losing water and organelles in the process The loss of nuclei is seen in the bow region These processes lead to gradual increases in protein concentration, refractive index and hardness in the central areas Consequently, there are a considerable number of post-translational molecular changes that take place in the lens throughout life With aging, the lens often loses some of its transparency When transparency is reduced to the point where visual acuity is impaired, the condition of cataract exists Cataract is also associated with a number of other clinical conditions (such as diabetes) and

Figure 3 – 2 Diagrammatic section of the human eye

Figure 3- 3 Diagrammatic representation of the mammalian lens environmental stresses (such as radiation).

Bibliographical study at calcification 62

Clinical reports on IOL calcifications

Calcification of medical devices fabricated from polyurethane, silicones, and hydrogels has been widely reported These devices include bioprosthetic heart valves, cerebrospinal fluid shunts, mammary implants, nose implants, artificial finger joints, and contact lenses (Figure 3-4) 10-16 Calcifications of ophthalmic devices have also been increasingly observed following implantation 17-24

The first case of calcification seems to have appeared from the setting of the 1980s during the implantation of the IOLs in the posterior chamber Medicines and Healthcare products Regulatory Agency (MHRA) issued device alerts to all UK hospitals about an IOL packaging error, which can cause cloudy vision Two specific brands are involved, Hydroview TM H60M (Bausch and Lomb, 1997-2001) and Aqua-SenseTM (Ophthalmic Innovations International Inc, 1999-2000) 25

By comparing clinical and pathological features of explanted hydrophilic acrylic intraocular lenses of three major designs, Hydroview TM , SC60B-OUV TM and Aqua-Sense TM , Dr Apple group found that 1) the irregular granular deposits were detected on the external optical surfaces of Hydroview TM lenses; 2) with the SC60B-OUVTM lenses, the opacity was caused by the presence of multiple fine, granular deposits within the lens optic; 3) the Aqua-Sense TM lenses exhibited both patterns simultaneously 26 They suggested that differences in the water content of the hydrophilic acrylic materials (Hydroview TM , 18%; SC60B-OUV TM , 28%; and Aqua-Sense TM , 25%) used in the

Figure 3-4 Photographs taken under an operating microscope immediately after explantation of the SC60B-OUV lenses (Medical Developmental Research, Inc., Clearwater, FL) from a 64-year-old diabetic female The surgeon noted that the optic of the lenses actually resembled a nuclear cataract (Case of Dr Mahmut Kaskaloglu, Alsancak Izmir, Turkey.) (A) Note the complete opacification of the intraocular lens (IOL) optic and haptics in this lens This lens was explanted and exchanged from the left eye 21 months after implantation (B) The opacification was limited to the IOL optic in this lens explanted from the right eye of the same patient 11 months after implantation calcium precipitation

A German medical group 27 reported that patient-related factors, systemic disease (e g, diabetes mellitus) and ocular inflammation, may contribute to IOL calcification Dr Joo and his coworker also indicated that for diabetic patients the breakdown of the blood- aqueous barrier is a possible causative factor in early inflammation after surgery 28

According to the hypotheses collected in these articles, certain factors seem to play an important role in the process of calcification or crystallization:

Adjuvant products in ophthalmologic surgery

Viscoelastic substances (OCUcoat/ Viscoat/ Amvisc Plus, Table 3-1) are indicated for use as surgical aids in ophthalmic anterior 29 and posterior 30 segment procedures including cataract extractions, intraocular lens (IOL) implantations, corneal transplantation surgery, glaucoma filtering and retinal reattachment procedures Some calcification phenomena have been reported to occur on the IOL surfaces intraoperatively during cataract surgery with implantation of some silicone IOLs 18, 30 or in the early postoperative period after implantation of some hydrogel IOLs However, significant depositions of crystalline materials on IOL surfaces are uncommon Previous studies have suggested that the deposits were composed of calcium phosphates but the mechanism is not fully understood 17, 31, 32 The question arises as to whether this mineral

Previous studies have shown that traces of long chain saturated fatty acids are present in ocular fluids, 34 and that the levels of these free acids was significant in senile cataracts 35 Fatty acids that have been identified in the aqueous humor include myristic, palmitic, stearic, arachidic and behenic Considerable research has been directed toward elucidating the mechanism of long-chain fatty acid transport across cell membranes 36 It was found that the adsorption of the fatty acid molecule is chain-length-dependent 37 However, there are no reports concerning calcification that may take place at the interface between adsorbed fatty acids and intraocular lens surfaces

Octacalcium phosphate [OCP, Ca8H2(PO4)6 5H2O], which has been proposed as a precursor phase in the formation of many biological apatites, prior to transformation to the thermodynamically more stable HAP These precursors are more soluble and have rate constants for crystal growth considerably greater than that for HAP in aqueous supersaturated calcium phosphate solutions The presence of an apatite-like layer in OCP may allow for the epitiaxial over-growth of HAP and OCP 38-41 OCP also appears to play a significant role in the chemistry of bones, teeth, phosphate fertilizers, and other precipitated calcium phosphates.

Calcification Study in vitro 67

Hydrogel foldable lenses (Ultem and Surefold gasket), intraocular lenses (Hydroview TM , H60M), optical implants for the replacement of human crystalline lenses, and viscoelastic substances, OCUcoat (LOT 1356), Viscoat (LOT 01E10A) and Amvisc Plus (LOT B010420)), were provided by Bausch & Lomb Myristic acid (LOT 99H0725), palmitic acid (LOT 110K03011), stearic acid (LOT 49F8449), arachidic acid (LOT 41K1558) and behenic acid (LOT 96F84151) were from SIGMA Chemical Co

Stock solutions of calcium chloride (CaCl2), sodium chloride (NaCl), potassium dihydrogen phosphate (KH2PO4), were prepared in calibrated grade A volumetric glassware using triply distilled de-ionized water (TDW, made in our laboratory) and reagent grade chemicals dried under vacuum at 110 o C Carbon dioxide-free potassium hydroxide solution (KOH) was prepared in a nitrogen atmosphere from washed Reagent Grade pellets All solutions were vacuum filtered twice (0.22àm Millipore filters) before use and analyzed

Calcium ion solution concentrations were determined by complexometric titration with disodium ethylenediamine-tetra acetic acid, EDTA (J T Baker), and Eriochrome Black-T as an indicator Potassium dihydrogen phosphate solution concentrations were confirmed using UV-Vis spectrophotometric analysis for phosphate as the phosphor- vanadomolybdate yellow complex Potassium hydroxide solutions were titrated against a standardized hydrochloric acid solution, DILUT-IT (J T Baker), with phenolphthalein as indicator

The lenses had been lathe cut and polished from the central hydrogel portions of composites, which contained a bonded UV absorber The haptics were formed from the outer polymethylmethacrylate (PMMA) portion

To accelerate the possible interactions between IOLs and the gasket and holder materials of the vials, the polypropylene holder/folders (15 units, Lot 12483) and retainers (25 units, Lot 12489) were Soxhlet-extracted with petroleum ether (AR grade, bp 35 - 60 o C) for 48 hours The solvent was removed in vacuo affording 0.46 g of pale yellow oil A solution of 9.4 mg of the extract was diluted to a total weight of 9.40 g with HPLC grade hexane A total of 18 freshly manufactured, Ultem-packaged IOL’s ( H60M Lot 44VD) were air-dried on a sheet of PTFE overnight Nine of them were dipped into this diluted extract prior to air drying Each IOL was then replaced in its vial with the original water

Constant composition (CC) 42 OCP crystallization experiments on the IOL surfaces were made in double-walled Pyrex glass vessels maintained at 37.0 ± 0.05 o C Supersaturated reaction solutions (σOCP=2.28; where the relative supersaturation σ, is defined as (IAP/Ksp) 1/16 –1, IAP and Ksp being the ionic activity and solubility products of OCP, respectively) were prepared by the slow mixing of calcium chloride (2.50x10 -3 mol L -1 ) and potassium dihydrogen phosphate (1.88x 10 -3 mol L -1 ) with the ionic strength maintained at 0.15 mol L -1 by the addition of sodium chloride solution A 0.06 mol L -1 of potassium hydroxide solution was added over a period of 45 minutes to the supersaturated solution to adjust the pH to 7.10 ± 0.05 A pH electrode (Orion 91-01) coupled with a single-junction reference electrode (Orion 90-01) along with a pH meter presaturated with water vapor, was bubbled continuously through the reaction solutions during pH adjustment and crystallization experiments

Once effective equilibrium of the metastable supersaturated solutions had been attained, the IOL samples were introduced, and the compositions of the reaction solutions were maintained constant by the simultaneous addition of two titrant solutions from mechanically coupled burets, one containing calcium and sodium chlorides, and the other potassium dihydrogen phosphate and potassium hydroxide The concentrations of the titrant solutions were calculated using Eqs (3-1) - (3-4)

[CaCl2]t = 2[CaCl2]rs + 4Ceff (3-1) [NaCl]t = 2[NaCl]rs - 8Ceff (3-2) [KH2PO4]t = 2[KH2PO4]rs + 3Ceff (3-3) [KOH]t = 2[KOH]rs + 5Ceff (3-4)

In Eqs (3-1) - (3-4), the subscripts t and rs refer to the concentrations of titrants and metastable supersaturated reaction solutions, respectively Ceff is the effective concentration of added titrants with respect to OCP, or the moles of OCP formed per liter of mixed titrants The value chosen for Ceff for this crystallization study was 2.00×10 -4 mol L -1

During the experiments, the constancy of concentrations was verified by analysis of filtered aliquots (0.22àm Millipore filters) for calcium by atomic absorption spectrometry (Perkin Elmer Atomic Absorption Spectrometer 3100) and for phosphate spectrophotometrically as the vanadomolybdate complex (Hewlett-Packard, 8452A, Diode Array Spectrophotometer) In all cases, the concentrations remained constant throughout the precipitation reactions to within ±1.5%

After washing with triple distilled water, the IOL samples were air-dried overnight and examined using Scanning Electron Microscopy (SEM; JEOL JSM-5300, Noran Instrumental Inc Middleton, WI) and Diffuse Reflectance Infrared Fourier Transform Spectroscopy (FTIR, Perkin Elmer 1760S FT-IR spectrometer)

The solubility product of the OCP used in these calculations was taken as 2.51 x

10 -99 (mol L -1 ) 16, 43 Ionic concentrations were computed from mass balance, proton dissociation, electroneutrality, and equilibrium expressions involving phosphate protonation and ion-pair formation with the calcium ion as described previously 44, 45

Surfaces were characterized by Dr George Grobe (Bausch & Lomb) using a Phymetrics (PHI) Quantum 2000 Scanning XPS Microprobe incorporating a monochromatic Al anode operated at 15kV and 40 Watts in standard power mode with dual beam neutralization (ions and electrons) Static point acquisitions were collected with a 200 àm 2 analysis area The base pressure of the instrument was 5 x 10 -10 torr and during operation the pressure was ≤1 x 10 -7 torr This instrument made use of a hemispherical analyzer operated in FAT mode A gauze lens was coupled to a hemispherical analyzer in order to increase signal throughput Data were collected with a SUN Ultra 5 workstation running Solaris 2.6 OS and COMPASS v 4.0A The instrument utilized MultiPak version 6.0 software for data analysis Scanned X-ray images (SXI) were collected using a focused x-ray beam, which causes the emission, from the specimen surface, of photoelectrons, Auger electrons and secondary electrons in competing processes The analyzer was set (stage bias, gauze lens to +90 volts) to accelerate the secondary electrons into the analyzer for imaging The secondary electrons two-dimensional image of the specimen This image was then used to define the area for X-ray potoelectron spectroscopy (XPS) analysis

For XPS examination, the IOL lenses were placed on a stainless steel mount and held in place with a molybdenum mask Lens surfaces were analyzed utilizing low- resolution survey spectra (0-1100eV) to identify the elements present Quantification of elemental compositions was completed by integration of the photoelectron peak areas Analyzer transmission, photoelectron cross-sections and source angle correction were taken into consideration in order to give accurate atomic concentration values

OCP crystallization on Hydroview IOL surfaces

Freshly manufactured, Hydroview H60M IOLs were tested both uncoated, and coated with OCUcoat, Viscoat, or Amvisc Plus (see Table 3-2) The lenses were placed in inserters (IC-2BU, Model IM002) containing a few drops of the Viscoelastic substances and the coated lenses were delivered by injection, into 2.0ml of 2.94×10 -4 mol L -1 behenic acid in 90% ethanol solution and equilibrated for 30 minutes The value chosen for Ceff for this crystallization study was 1.00×10 -4 mol L -1

A total of 15 Hydroview H60M IOLs, air-dried overnight, were coated with 2àl of 6.44×10 -3 mol L -1 cyclic silicone (octamethylcyclo-tetrasiloxane) in hexane, air dried (about 2 hrs), and immersed for 30 min in 2ml of the fatty acid solutions (myristic acid /palmitic acid /stearic acid /arachidic acid /behenic acid) at concentrations ranging from 4.00×10 -5 mol L -1 to 2.94×10 -4 mol L -1 (Table 3-3) They were rinsed in triple distilled

Table 3-2 Crystallization tests of OCP on Hydroview IOLs Coated with Different Viscoelastics

Exp.No Viscoelastics Treated with

1 None (control) 30min 650 no no Clear

5 None (control) no 630 no no Clear

9 OCUCOAT 30min 720 no no Clear

13 OCUCOAT no 670 no no Clear

16 VISCOAT 30min 310 no no Clear

19 VISCOAT no 300 no no Clear

23 VISCOAT 30min 240 yes yes Visible

28 AMVISC PLUS 30min 60 yes yes Visible

31 AMVISC PLUS no 70 no no Clear

32 AMVISC PLUS no 50 no no Clear

Table 3-3 Crystallization tests of OCP on Hydroview IOLs treated by cyclic silicone and fatty acids

S9 Stearic (0.4) 400 water before inserting into the supersaturated solutions

Nucleation Tests on Hydrogel IOL surfaces

Experimental conditions and induction times, τin, preceding nucleation, are summarized in Table 3-4 The reaction solution was highly supersaturated with respect to OCP (σOCP = 2.28) but stable for more than 10 hours, after which OCP crystals precipitated on the walls of the reaction cell and/or the electrode A typical titrant volume curve as a function of time for OCP nucleation on an IOL surface is shown in Figure 3-5 The induction time, τin, needed to reach steady-state nucleation, was measured as the time from the introduction of IOL surfaces to the intersection of a volume/time tangent (dV/dt) with the time axis Although there is no precise method to draw the tangent, for consistency it was obtained from the linear regression of titrant addition (0.5 - 1.5 mL) volumes as a function time (see Figure 3-6) The reproducibility of the induction periods depends both upon their magnitude and the thermodynamic driving force In all the tests three stages are clearly distinguished (Figure 3-6) First, in the induction region, the solution concentrations and pH were unchanged and the volume, dV, of added titrant remained essentially zero confirming the absence of mineral nucleation or growth In second stage, titrant addition commenced as heterogeneous nucleation set in This stage reflected the commencement of stable nucleus formation with an approximately linear titrant addition; this line was extrapolated to zero titrant to estimate the induction time In the third region, more rapid addition reflected both heterogeneous nucleation and the exponential growth of OCP crystals on the nuclei formed in the second stage

Figure 3-5: Plot of titrant addition as a function of time for calcification test of OCP on IOL lens

(Lot No 44VD2) packaged in Ultem vial after treatment with behenic acid

10 ov er growt h het erogeneous nucleation induct ion r egion

Figure 3-6: Plot of titrant addition as a function of time for calcification test of OCP on

Inhibition of Dicalcium Phosphate Dihydrate Crystallization by Additives 100

Inhibition of Dicalcium Phosphate Dihydrate Crystallization by Magnesium Ions 102

Many studies have been made of the influence of magnesium on the precipitation of calcium phosphates Brown and Bachra suggested that in vivo, magnesium lengthens the lifetime of amorphous calcium phosphate (ACP) and octacalcium phosphate (OCP) 15, 16 Boskey and Posner further proposed that magnesium reduces the solubility of ACP, thereby increasing the induction period for the transformation to HAP 17 Newesely also reported that the Mg 2+ cation stabilizes tricalcium phosphate (TCP) in the precipitation of HAP 18

Several attempts have been made to determine the ability of magnesium to incorporate into the apatite lattice It has been well established that magnesium can

Figure 4-1 12 Crystal structure of DCPD as viewed down the b-axis

However, several groups believed that magnesium may be substituted into the apatite crystals to a limited extent even though the ionic radius of magnesium is smaller (0.65Å) than that of Ca (0.99Å) 22-24 For instance, Featherstone and co-workers have demonstrated that a carbonate-apatite structure can incorporate magnesium in the apatite lattice, whereas pure hydroxyapatite will exclude it They were also able to show that the crystallinity of carbonated apatites formed in the presence of magnesium was improved 25 Since magnesium is a minor constituent of tooth enamel (about 0.6% by weight), Terpstra and Driessens proposed three possibilities for its presence in tooth enamel: (1) magnesium is incorporated into the apatite lattice; (2) it is surface bound; or (3) it is present as a separate mineral phase 20

At this time no definitive evidence exists which associates the action of magnesium with the incidence of caries; these are conflicting results regarding in the literature Thus although it has been reported that high levels of magnesium in water supplies are associated with low caries prevalence, 26, 27 it has also been reported that during the early stages of enamel caries, magnesium may participate and cause preferential dissolution in one of three areas: (1) surfaces rich in magnesium; (2) apatite crystals containing magnesium throughout the structure; or (3) a separate calcium phosphate phase, rich in magnesium 28, 29

Nancollas suggested that the influence of foreign ions on crystal growth may occur by (a) additives complexing with lattice ions and preventing them from participating in the crystal growth reaction; (b) increasing the ionic strength of the working solution thereby increasing the solubility of calcium phosphate phases; or (c) adsorbing onto

While the other calcium phosphate phases are known to incorporate magnesium, there has been no evidence that brushite does so To gain a better understanding of how magnesium interacts with brushite, CC experiments were performed over a magnesium concentration range from 8.5×10 -5 mol L -1 to 2.55×10 -3 mol L -1 In addition, AFM, SEM and XRD measurements in the presence of magnesium ions were also made

DCPD seed crystals were synthesized as described in Chapter 2 Dried powders were used in DCPD growth experiments Supersaturated solutions were prepared as discussed in Chapter 2, using CaCl2, KH2PO4, and MgCl2 The ionic strength of each solution was adjusted to 0.15 mol L -1 with NaCl and the temperature was maintained at 37.0 ± 0.1 o C

Table 4-1 lists the experimental conditions for DCPD growth in the absence / presence of magnesium ions DCPD supersaturated solutions were prepared at a pH of 6.00, and σ DCPD = 0.236 Various concentrations of 8.50 ×10 -5 molL -1 - 2.55×10 -3 molL -1 magnesium ion were introduced to the reaction solution prior to adjusting the pH

The pH of DCPD experimental solutions was adjusted by the slow addition of KOH The volume of the KOH solution was accounted for when calculating the total reaction solution volume Experiments, initiated through the addition of dry DCPD seed crystals, were controlled by the addition of titrant solutions, whose concentrations were calculated from the equation given in Table 2-3 The additions were triggered by a potentiometer (Orion 720A), incorporating a glass pH electrode (Orion 91-01) and a reference electrode (Orion 900100) Aliquots of the reaction solutions were withdrawn periodically during the experiments to verify constant composition

Table 4-1 Reaction conditions for DCPD crystal growth in the absence/presence of Mg 2+

Growth rates were determined from plots of the volume of added titrants as a function of time (Fig.4-1-1) The overall growth rate, R is defined by Eq.(2-12) in Chapter 2 seed t eff m SSA dt C dV

=( )⋅ where dV t dt is the rate of titrant addition (L min -1 ), C eff is the number of moles of deposited crystal per liter of added titrant, m s is the seed mass (mg), and SSA is the effective surface area (m 2 )

The % inhibition values were calculated using Eq 4-1

R Inhibition R i (4-1) where R 0 and R i are the rates in the absence and presence of the additive, respectively

Typical CC growth curves at σ = 0.236 in the absence and presence of magnesium are shown in Fig.4-1-1 Increasing the concentration of magnesium decreased the crystal growth rates derived from measurements of added titrant volumes versus time It can be seen that the relative reaction rate was significantly decreased at magnesium concentrations of 2.55× 10 -3 mol L -1 (Fig.4-1-2) Interestingly, at a magnesium level of 8.50 × 10 -5 to 2.83 × 10 -4 mol L -1 , the initial rate of DCPD growth was the same as that in the absence of additive at σ = 0.236 and I = 0.5 mol L -1 (Control) However, the presence of 8.50 mM magnesium caused the DCPD growth rate in pure supersaturated

Figure 4-1-1 Plots of titrant volume against time for DCPD crystal growth in the absence and presence of magnesium (σ = 0.236, pH = 6.0, IS = 0.15 mol L -1 , Ceff = 3.0×10 -3 mol

Control [Mg]=0.91 nM [Mg]=3.64 nM [Mg]=7.27 nM [Mg].91 nM

Figure 4-1-2 Inhibition of DCPD crystal growth in the presence of magnesium R is the

DCPD crystal growth rate (σ = 0.236, pH = 6.0, IS = 0.15 mol L -1 , Ceff = 3.0×10 -3 molL - , 12.6 mg seed added, 37 o C) supersaturated solution, 3.44 × 10 -6 mol m -2 min -1 (n=6) to decrease to 3.26 × 10 -6 mol m -2 min -1 (n=7) with a further progressive decrease to 1.60 × 10 -6 mol m -2 min -1 (n=5) at 2.55 mM magnesium (53.5.% inhibition) was reduced to about one third of that in control

No detectable changes of DCPD crystal morphology were found on the crystals after 6 hrs of growth in the presence of magnesium, see Fig 4-1-3

Evans and Johansson suggested that the inhibitory properties of Mg are mainly related to the growth of calcium phosphate phases 32, 33 It should also be noted that a reduced Ca / Mg ratio retards the formation of DCPD Although the risk of stone formation probably cannot be eliminated by such a shift in urine composition, the formation of a calcium phosphate crystal phase other than DCPD is favourable because of its high recurrence risk and shock wave-resistant structure

Typical plate-like brushite crystals with smooth (010) surfaces served as suitable substrates for AFM (the Miller indices (hkl) denote a single plane and [uvw] specify a unique vector direction) Parallel in situ AFM experiments, made under the same conditions as those for the CC experiments, show that brushite crystals in pure solutions grow on atomic steps generated at complex dislocation hillocks This behavior was also observed for other solution-based crystals growing near equilibrium 9, 33-36 These are triangular in shape with crystallographically distinct steps along the [101], [201], and [001] directions (Figure 4-1-4a) In the absence of magnesium, AFM images demonstrate that the DCPD growth hillock morphology is characterized by well-formed and straight step-edges in magnesium-free solution (σ = 0.236, [Mg] / [Ca] = 0), see Figure 4-1-4a They display anisotropic spreading velocities in the order [101]>[001]≈[201], which is

Figure 4-1-3 SEM and Energy-Dispersive Spectroscopy results for DCPD crystal growth in the absence/presence of magnesium.(a) [Mg] = 0 molL -1 ; (b, c) [Mg] = 0.00255 molL -1

Figure 4-1-4 Atomic Force Microscopy images demonstrate the effect of Mg 2+ on DCPD growth hillock morphology [Mg]/[Ca] (a)0; (b)1/10; (c) 1/3; (d) 1/1; (e) 3/1 a b c d e

The addition of Mg 2+ to the DCPD growth solutions yielded steady-state hillock morphologies that varied as a function of Mg 2+ concentration (Figs.4-1-4 b - e) Small additions of Mg 2+ to the growth solutions (low Mg/Ca ratios) had no effect on the steps However, in the solution with high Mg/Ca ratios, Mg 2+ stabilizes the [001] step at the expense of the [201] step as observed by the change in the relative length of the steps and the change in corner position Similarly, the velocities of [201] step are constant in the presence of low Mg 2+ concentration High Mg 2+ impurities change the [201] step spreading speed, see Figure 4-1-5

Modulation of Dicalcium Phosphate Dihydrate Crystallization by Osteopontin 129

Bone contains an extensive mineralized matrix of apatite crystals as well as a number of proteins including phosphoproteins and osteonectin 44, 45 Osteopontin (OPN), one of the main phosphoproteins in bone, is widely believed to be involved in remineralization due to its participation in the binding of osteoclast cells to bone 46 Since OPN is produced by a variety of cells, it can be found in many different tissues In the kidney, OPN is found in the loop of Henle and the distal convoluted tubules In bone, it is found within the interfibrillar compartment of the organic matrix and in the laminae limitantes This protein is also found in many biological fluids, such as blood, urine and seminal fluid 46, 47

Osteopontin is composed of an amino acid chain of ~300 residues, of which 25% are acidic, 23% are hydrophobic, 9% are basic, and 18% are serine residues Within this protein, the RGDS (arginine, glycine, aspartate, serine) motif attracted much interests because it is located in the mid-portion of the molecule and in a thrombin cleavage site just carboxyl-terminal to it Carbohydrates and sialic acids constitute 16.6% and 7% of osteopontin molecule, respectively 48, 49

The complete amino acid sequences of OPN are known for a variety of mammals and are extensively conserved, with 50% being identical and ~10% conserved The amino- and carboxyl-terminal regions, the RGDS motif and the aspartic acid sequence are all highly conserved, as well as the thrombin cleavage site and several serine phosphorylation sites The amino acid sequence of rat osteopontin protein was provided by Dr J.R Hoyer from the University of Delaware (Fig 4-3-1) 51-53

Figure 4-3-1 10 Osteopontin sequences most involved in minerallization

OPN is an important macromolecule involved in the inhibition of various steps of crystal formation It is also a major component of renal stones raising the question whether it acts as inhibitor or promoter during stone formation 54, 55

Rat bone osteopontin (molecular weight, 33,000) was isolated by immunoaffinity purification as described by Hoyer et al 56 Aliquots of OPN solutions were stored in sterile centrifuge tubes at -80 o C prior to use OPN was reconstituted by adding triply distilled deionized water DCPD seed crystals were synthesized and characterized by XRD and chemical analysis, as described in Chapter 2

Supersaturated solutions were prepared, as described in Chapter 2, using CaCl2 and

KH2PO4 The ionic strength of the reaction solutions was adjusted to 0.15 mol L -1 with NaCl and the reaction temperature was maintained at 37.0 ± 0.1 o C The pH was adjusted with standard KOH solution, whose volume was accounted for when calculating the total reaction solution volume The addition of titrant solutions were triggered by a potentiometer (Orion 720A), incorporating a glass pH electrode (Orion 91-01) and a Ag/AgCl reference electrode (Orion 900100) Aliquots of the reaction solutions were withdrawn periodically during the experiments to verify constancy of composition

Table 4-3-1 lists the experimental conditions for DCPD growth in the absence / presence of osteopontin DCPD supersaturated solutions were prepared at a pH of 5.6, and σ DCPD of 0.141 and 0.250 Various concentrations 0.03 àg ml -1 – 0.36 àg ml -1 of OPN were introduced to the reaction solution prior to the initial adjustment of pH (pH 5.6)

Table 4-3-1 Reaction conditions for DCPD crystal growth in the absence / presence of osteopontin σ = 0.141 σ = 0.25

Total volume 200.0 ml 100.0 ml 200.0 mL 100.0 ml

Ionic strength 0.15 mol L -1 0.15 mol L -1 pH 5.60 5.60

Figure 4-3-2 shows typical CC curves for DCPD growth in the presence of OPN (pH=5.6 and ionic strength is 0.15 mol L -1 ) The DCPD growth rates (σ = 0.141), R(mol m -2 min -1 ), at OPN concentration of 0, 0.91, 3.64, 7.27, and 10.91 nM respectively, are (2.49 ± 0.08) × 10 -7 (n = 7), (1.51 ± 0.04) × 10 -7 (n = 4), (8.31 ± 0.05) × 10 -8 (n = 5), (5.54 ± 0.04) × 10 -8 (n = 3), and (2.99 ± 0.05) × 10 -8 (n = 5), respectively (Fig 4-3-3) It can be seen that the rates gradually decreased with an increase in the citrate concentration, about 40% inhibition of growth in the presence of 0.91nM OPN, eventually approaching a maximum degree of inhibition of about 90% at an OPN concentration of 10.91 nM (Table 4-3-2) Therefore, OPN is an effective DCPD crystal growth retardant

Figure 4-3-3 and 4-3-4 are suggestive of Langmuir adsorption of OPN at the crystal surface This model assumes reversible additive adsorption at a finite number of identical sites on the surface At equilibrium, the rates of adsorption (J ads ) and desorption (J des ) are equal Therefore, the surface coverage, θ c can be defined by Eq 4-8 des ads J

Where k ads and k des are the specific rate constants for adsorption and desorption per unit concentration, C, respectively Γ m is the maximum value for ‘monolayer’ coverage, and Γ i , the moles of inhibitor molecules per unit surface Considering Eq 4-9

[OPN]=0.91 nM [OPN]=3.64 nM [OPN]=7.27 nM [OPN].91 nM

Figure 4-3-2 CC growth curves of DCPD in the absence / presenceat of different concentration of OPN (the curves have been normalized to the same seed mass of 22.1 mg, σ = 0.141)

Figure 4-3-3 DCPD crystal growth rates in the presence of various concentrations of

OPN Data points represent the average of five measurements (σ=0.141, pH=5.60)

Figure 4-3-4 Langmiur – type plot of DCPD growth in the presence of OPN (σ = 0.141, pH = 5.60)

The measured growth rates as a function of adsorbent concentration can be used to test the Langmuir model by Eq 4-12 57

K = is the affinity of the inhibitor for the surface and b is a limiting factor accounting for the effectiveness of the additive If b = 0, the adsorbate is capable of completely inhibiting the rates of growth at concentrations approaching infinity For

Langmuir-type adsorption behavior, a plot of R 0 (R 0 R i )

− as a function of 1 should C yield a linear plot with y-intercept 1/(1−b) equal to 1, resulting in complete inhibition at infinite inhibitor concentrations Figure 4-3-4 shows the linear fit of a plot of

− as a function of 1 [ OPN ] has a y-intercept of 1.08, strongly suggesting a Langmuir type adsorption fit

Similarly, at higher supersaturation (σ = 0.25), the DCPD growth rates R (mol m -2 min -1 ), are (8.21 ± 0.04) × 10 -7 (n = 7), (5.08 ± 0.05) × 10 -7 (n = 4), (1.98 ± 0.06) × 10 -7 (n = 4), (1.35 ± 0.06) × 10 -7 (n = 4), and (1.11 ± 0.07) × 10 -7 (n = 5), respectively (Figs

4-3-5 and 4-3-6) Figure 4-3-7 shows the linear fit of a plot of R 0 (R 0 R i )

− as a function of 1 [ OPN ] has a y-intercept of 0.97, which strongly suggests a Langmuir type adsorption fit

Influence of OPN on DCPD Morphology through Binding to Atomic Steps

AFM images show that growth of DCPD crystals (σ = 0.141) occurs on atomic steps

Conrol [OPN] = 6 à gml -1 [OPN] = 24 à gml -1 [OPN] = 48 à gml -1 [OPN] = 72 à gml -1

Figure 4-3-5 CC growth curves of DCPD in the absence / presenceat of different concentration of OPN (the curves have been normalized to the same seed mass of 22.1 mg, σ = 0.250)

Figure 4-3-6 DCPD crystal growth rates in the presence of various concentrations of

OPN Data points represent the average of five measurements (σ =0.250, pH =5.60)

Figure 4-3-7 Langmiur – type plot of DCPD growth in the presence of OPN (σ = 0.250, pH=5.60)

OPN brushite grows without significant morphological changes Steps grow underneath the majority of clusters and around the minority (see marked areas in Figs 4-3-8 a and d) However, steps grow until they are heavily pinned and no new steps are sourced in the presence of 10.91 nM OPN The “banded” distribution of OPN clusters (Fig 4-3-8 d) may indicate initial deposition of clusters on terraces near step edges or OPN clusters adsorb preferentially to step edges forming a denuded zone on the adjacent terrace In the high concentration of OPN growth are fully stopped by pinning Osteopontin inhibits brushite growth at higher supersturations (σ = 0.25) in similar way as it does at lower supersaturations, see Fig 4-3-9

In the presence of OPN, although the grown brushite crystallites were much smaller due to the rate reduction, they retained their plate-like shape, see Fig 4-3-10 This suggests that the adsorption of OPN does not occur only at specific faces, rather, these macro molecules are able to interact with the dominant (010) face and retard its growth much more effectively Thus, prominent brushite faces are preserved in the presence of this additive In comparison of crystallites grown in a pure solution, the thickness of the crystals in Fig 4-3-10 was increased, confirming that the retarding influence of OPN was stronger on the (010) face, compared to lateral faces The strong phosphorylation of the OPN molecule enhances its anionic charge, and it is now well documented that aspartate residues on acid-rich proteins contribute to their strong interaction with crystal faces 53, 58 Although OPN is highly flexible, the binding events at the heterogeneous interface are inevitably complicated by such factors as the size of the protein as well as the local properties of the mineral face to which the protein attaches 4 The results would indicate that, in addition to strong electrostatic interactions, a structural and sterochemical fit

Table 4-3-2 Growth rates and % inhibition values of DCPD in the absence / presence of osteopontin σ = 0.141 σ = 0.250

Figure 4-3-8 AFM frames of a growing (010)-brushite face (σ = 0.141), (a, b, c) in the absence of OPN; (d, e, f) in the presence of OPN at 0.91, 3.64, and 10.91 nMrespectively Introduction of OPN molecules reduces the brushite growth rate by inhibiting step sourcing and causing step bunching at low concentration (e), and pinning steps at high concentration (f)

Figure 4-3-9 AFM frames of a growing (010)-brushite face (σ = 0.25), (a, b, c) in the absence of OPN; d, e, f)in the presence of OPN at 0.03, 0.12, and 0.36 àg ml -1 respectively Introduction of OPN molecules reduces the brushite growth rate by inhibiting step sourcing and causing ste bunching at low concentration (e), and pinning steps at high concentration (f)

Figure 4-3-10 SEM of grown brushite in the presence of OPN ([OPN] = 0.12 mg ml -1 , σ= 0.250, pH = 5.60) probably exists between the periodical spacing of the inhibitors

Referneces 147

An Understanding of Renal Stone Development under a Simulated Oxalate-

Studies of abnormal biomineralization processes in urinary stone formation has provoked interest by scientists in medical, chemical, material and biological fields [1] It has been shown that at least 50% of calcium oxalate stones contain variable amounts of apatite, occasionally in a “nuclear” location [2-8] It has also been suggested that calcium phosphate (CaP) may nucleate and govern the development of calcium oxalate calculi [9] However, the mechanisms of formation of renal calcium oxalate containing concrements are still not fully understood

There is evidence that Randall’s plaques seem to start as a precipitation of calcium phosphate, notably brushite (DCPD, CaHPO4.2H2O), either in the loop of Henle or in the distal part of the distal tubule [10-14] It is also postulated that brushite could induce the heterogeneous crystallization of CaOx [15-19] and that supersaturation of brushite may be the fundamental abnormality in calcium oxalate stone initiation [20] Tiselius found that crystals of CaP that form high in the nephron might be dissolved when they are exposed to the acidic urine in the collecting duct [21, 22] Akbarieh and Tawashi reported that surface transformation of calcium oxalate trihydrate (COT) calcium oxalate monohydrate (COM) occurred only in stone formers urines [23] Therefore, a better understanding of CaOx and CaP crystallization, phase transformation, and their participation in progression to stones is timely

In the present work, one particularly useful technique, Dual Constant Composition (DCC), was used because it can maintain, in vitro, the supersaturation levels normally found in different areas of the urinary tract [24] It also provides a well-controlled chemical system by which chemists can study the effect of thermodynamic

An Understanding of Renal Stone Development in a Simulated Oxalate-Phosphate System 151

Renal stone formation 153

In the industrialized world, at least 10% of the population is afflicted by urinary tract stone disease Urolithiasis occurs in both men and women but the risk is generally higher in men In Northern Europe and North America, more than 75% of the stones are composed of calcium oxalate with or without an admixture of calcium phosphate Approximately 5% of the stones are composed of pure calcium phosphate, 5–15% of uric acid/urate, 10% of infection stone material, such as struvite and carbonate apatite (CAP) and 1 % of cystine 25-27

The crystalline components frequently found in urinary calculi are shown in Table 5-1 28 Their precipitation as stones is presumably governed by the degree of urine supersaturation with respect to each of the mineral phases Frequently, the laminar structure of stones is the result of oriented overgrowth of one crystalline phase upon another, reflecting the variations in urine chemical composition during the growth of individual calculi

In stones of mixed composition, a knowledge of the depositional sequence of mineral constituents is essential, in the nidus composition reflects the physical chemistry of the stone initiating process The other compounds present are the result of secondary

Table 5-1 Crystalline components of Urinary Calculi

Hydroxyapatite, HAP - Ca10(PO4)6(OH)2

Carbonate-apatite, CAP - Ca10(PO4, CO3, OH)6(OH)2

Calcium hydrogen phosphate dihydrate, DCPD

Tricalcium phosphate, TCP Whitlockite Ca3(PO4)2

Octacalcium phosphate, OCP - Ca4H(PO4)3 2.5H2O Magnesium hydrogen phosphate hexahydrate

Uric acid dihydrate - C5H4N4O3 2H2O Ammonium acid urate - C5H3N4O3NH4

- C5H3N4O3Na H2O Cystine - (-SCH2CHNH2COOH)2

Xanthine - C5H4N4O2 nucleation, growth, and aggregation of crystalline material following the initial formation of stone nucleus

The many reactions that may be involved in the formation of crystals in the urinary tract are summarized in Figure 5-1 29 Crystal nucleation, crystal growth, crystal aggregation, and crystal retention are considered to be the fundamental steps in stone formation

Whether the initial crystallization takes place as free or fixed particles it has widely been accepted that the precipitation of calcium oxalate is too slow to result in crystals of sufficient size to be trapped in the tubular system 30 Previous data indicated that free particles of calcium phosphate as well as of calcium oxalate might form at the levels of supersaturation that are occasionally reached in nephron urine 31 In this way crystals might become large enough to be trapped intratubularly

Figure 5-2 hypothetically interprets the possible events of normal crystallization in urine Under normal conditions, calcium oxalate and calcium phosphate crystals are small and well protected from further crystal growth and aggregation In this way, it is likely that small crystals can move through the tubular system and be excreted with urine It is also possible that small calcium phosphate crystals are completely dissolved during their transport through the collecting duct However, under appropriate conditions primary nucleation of calcium oxalate might occur in collecting duct urine As long as these crystals remain small and are well protected from growth and aggregation they exit the tubular system without problem 32, 33

Figure 5-2 36 Hypothetical interpretation of the possible series of crystallization events

(1) brush-border membrane of proximal tubular cells;

(2) repulsion between small calcium phosphate crystals;

(3) between calcium phosphate crystals and tubular cells;

(4) elimination of small calcium phosphate crystals by dissolution or passage with urine;

(5) internalization and intracellular dissolution of calcium phosphate crystals;

(6) primary nucleation of calcium oxalate;

(7) calcium oxalate nucleation induced by calcium phosphate;

(8) attachment of small calcium oxalate crystals to tubular cells;

(10) macrophage destruction of calcium oxalate crystals in the interstitial tissue, small intraluminal crystals of calcium oxalate are excreted with urine

Lieske and co-workers suggested that calcium oxalate monohydrate crystals adhere to tubular cells in a specific and rapid way 34 A year later they applied these same principles for crystal attachment crystals of hydroxyapatite (HAP) 35 The risk of crystal adherence is certainly greatest for calcium phosphate and calcium oxalate crystals that are large and thus move slowly through the nephron For very large crystals and crystal masses the repulsive forces described above are probably insufficient to counteract both crystal aggregtion and crystal – cell adherence

In pathological crystallization, low concentrations or structural abnormalities of crystallization modifying macromolecules may cause increased growth and aggregation of crystals, and large crystal masses of either of calcium phosphate or of calcium oxalate

37 Large crystal masses with or without sufficient protection by macromolecules will adhere to the tubular cell The crystals might alter the plasma membrane causing endocytosis to occur; whereas crystals of reasonable size can be taken care of and destroyed by the cell, larger crystal agglomerates might cause cell death 38, 39 It is understood that the insufficient or defect control of the crystallization process will result in development of large crystals or crystal agglomerates that remain within the tubular lumen Therefore calcium oxalate or calcium phosphate crystals might be trapped at the lower and narrow end of the collecting duct and thereby serve as a nidus for further crystal precipitation in the supersaturated urine

The various possibilities for calcium salt crystalluria and calcium stone formation are summarized in Figure 5-3 Small crystals of calcium oxalate or calcium phosphate that might have formed in the nephron can disappear either by intraluminal dissolution or

Fig 5-3 36 Possible outcomes of calcium salt crystallization in the urinary tract Small crystals of calcium phosphate or calcium oxalate can disappear by dissolution (1, 9) or remain small (3, 8) Nucleation of calcium oxalate might be induced by calcium phosphate (2, 6, 7) Following complete dissolution of calcium phosphate pure calcium oxalate aggregates can form (7) In constantly alkaline urine there is no precipitation of calcium oxalate and the result will be pure calcium phosphate aggregates, either in the form of hydroxyapatite (4) or brushite (5) A primary nucleation of calcium oxalate in acid collecting duct urine can lead to the aggregation of crystals containing calcium oxalate and calcium phosphate, or pure calcium oxalate Pure calcium oxalate aggregates might also form when calcium oxalate is primarily precipitated in the nephron unrelated to a calcium phosphate crystal phase 40 They also might form in those patients with a consistently high urine pH because calcium phosphate crystals do not dissolve.

Calcium oxalate 160

Although three hydrates of calcium oxalate (CaOx), monohydrate (CaC2O4 H2O, COM), dihydrate (CaC2O4 2H2O, COD) and trihydrate (CaC2O4 3H2O, COT), have been characterized, only COM and COD have been identified in urines Since urine is normally supersaturated with respect to all three phases, 41 the thermodynamically less stable COD and COT may transform to the more stable COM during development of renal stones

COM, or whewellite, is the thermodynamically most stable phase of CaOx It has a monoclinic structure, with a P21/c space group 42 The crystal habit is “coffin-like” as a single crystal but twinning is common, leading to an almost dumbbell-like shape, see Figure 5-4

COD, or weddelite, is often found naturally in kidney stones, plants, and fossils 43 It has tetragonal structure, with a space group of I4 and has a tent-lik or tetragonal bi- pyramidal crystal, see Figure 5-4

COT is the least stable CaOx phase and has a triclinic structure with a space group of P1 and a plate-like habit COT is well accepted as an important precursor to COM and COD, see Figure 5-4 44

Figure 5-4 44 The three calcium oxalate n-hydrate phases.

Experimental procedures 162

DCPD seed crystals were synthesized as described in Chapter 2 Table 5-2 lists the experimental conditions for the DCPD dissolution – COM crystallization experiments For DCC DCPD experiments, a metastable solutions, undersaturated (σ DCPD = -0.330) with respect to brushite and slightly supersaturated (σ COM = 0.328) with respect to COM, were prepared in magnetically stirred water jacketed Pyrex vessels at 37.0 o C with ionic strength, I = 0.15 mol L -1 , adjusted by the addition of sodium chloride Supersaturated solutions (150ml) were prepared by the introduction of filtered (0.22 àm Millipore filter) sodium chloride, calcium chloride, potassium dihydrogen phosphate solutions The pH was adjusted to 6.0 by the slow addition of potassium hydroxide solution Nitrogen, saturated with water vapor at 37 o C, was purged through the vessel in order to avoid contamination by atmospheric carbon dioxide The reactions were initiated by the introduction of known amount of dry DCPD seed crystals into the working solution The DCC apparatus, illustrated in chapter 2, where a calcium ion selective electrode (Orion) and a pH electrode (Orion), both coupled with an Ag/AgCl reference electrode, were used to control the addition DCPD dissolution titrant and COM growth titrant, respectively

For DCPD dissolution, a single titrant solution was used, which contained HCl and NaCl at the following concentrations

For the growth of COM, two titrant solutions were prepared, one containing KOH and

Table 5-2 Reaction conditions for DCPD crystal growth experiments

T KOH = 2W KOH , and T KH2PO 4 = 2W KH2PO4 (5-6) where W and T are the total concentrations in the reaction solutions and titrants, respectively, C eff,COM and C eff,DCPD are the effective titrant concentration with respect COM and brushite, respectively

During the experiments, solution samples were withdraw periodically, filtered (0.22àm Millipore filters) and analyzed for calcium by atomic absorption (Perkin Elmer, AAS 3100) and for phosphate spectrophotometrically (UV-Vis Monosorb ® , Quantachrome Instruments) as the phosphovanadomolybdate complex in order to verify the constancy of the solution composition (±1.5%)

The driving force for either the dissolution of DCPD in undersaturated solution or the crystal growth of COM in supersaturated solutions is controlled by the changes in Gibbs free energies The relative undersaturation of DCPD and the relative supersaturation of COM, are developed from Eq (1-1) in Chapter 1

The ionic activity products, IP, IP DCPD = + ⋅ 2 −

2 O a C are the activities of Ca 2+ ,HPO4 2- , and C2O4 2- ions The solubility activity products , KSP of DCPD and COM at 37.0 °C are 2.32×10 -7 mol 2 L -2 [45,

46] and 2.20 × 10 -9 mol 2 L -2 [47] , and respectively Speciation calculations were made using

Titrant buret #2 calcium and total phosphate with appropriate equilibrium constants by successive approximation for the ionic strength.

Results 165

5.4.1 CC Investigations of DCPD Dissolution

Figure 5-5a shows a typical plot of titrant volume as a function of time for DCPD dissolution at an undersaturation of -0.330 In agreement with previous studies, [50, 51] the dissolution rate decreased with time during the experiments It can be seen that the slope of the curve decreased markedly with time, eventually, tending towards zero with no detectable dissolution even though DCPD crystals remained in the undersaturated solution The XRD spectrum of the undissolved solid sample phases is shown in Figure 5-5c; the plate-like crystals (Figure 5-5b) were identified as brushite

5.4.2 CC Investigation of COM Nucleation

In practice, homogenous nucleation may not occur immediately after the creation of supersaturation Rather, a period of time, the “induction time”, τ, elapses before new crystallites persist in the supersaturated solution Although the induction period is a complex quantity, if the simplifying assumption is made that τ is essentially concerned only with the nucleation process, it can be described by Eq (5-8) [52, 53] ,

C γ SL τ ∝ + (5-8) where C 1 and C 2 are independent constants and γSL is the interfacial energy

Heterogeneous crystallization including epitaxial overgrowth of one crystalline phase upon another, can reduce the critical supersaturation and thereby, τ

Figure 5-5 (a) Typical CC plot of DCPD dissolution at σ = -0.326; (b) SEM image of

DCPD seed crystals; (c) X-ray diffraction pattern of DCPD seed crystals

Figure 5-6 (a) Typical CC plot of COM crystallization at σ = 0.376; (b) SEM image of

COM crystals; (c) X-ray diffraction pattern of COM crystals

In control, i.e unseeded nucleation experiments, the COM supersaturated solutions (σ = 0.328) were stable for 62 ± 2 (n = 4) hours (Figure 5-6a) Scanning electron micrographs of crystallites after crystallization showed the monoclinic COM morphology (Figure 5-6b)

5.4.3 DCC Investigation of DCPD Dissolution – COM Nucleation

Plots of titrant volume as a function of time (Figure 5-7) show that the brushite dissolution rate decreased and the addition of COM titrants reflecting nucleation and growth of this phase began only after an induction period of 300 ± 20 (n = 4) min The difference in induction time (τ = 62 hr in control experiment) indicates that the introduction of DCPD seed accelerates COM nucleation process Dissolution of DCPD increases the supersaturation with respect to calcium oxalate by increasing local concentration of calcium Nucleation of calcium oxalate can thus take place either by epitaxy on the surface of the dissolving DCPD crystal or by nucleation, with or without the contribution of a promoter, or by nucleation in the macromolecular environment that surrounds the DCPD crystals In Fig 5-5a, it can be seen that at the same undersaturation the slopes of the curves are quite similar, indicating that DCPD dissolution rates were not changed by the presence of oxalate

Fig 5-8 compares the X-ray diffraction patterns for the samples The DCPD seed was phase pure without the detection of other forms of calcium phosphate detected as shown in Fig 5-8a The DCC DCPD dissolution results in the presence of oxalate showed an unexpected calcium phosphate phase, monetite (CaHPO4) (Fig 5-8b) It is

Figure 5-7 DCC crystallization plots in a mixed solution of calcium phosphate (σbrushite

Figure 5-8 X-ray diffraction patters of DCC experiment (a) zero minute; (b) after 50 removal from the crystal lattice of the hydration water of brushite is one of the processes involved in brushite-monetite transformation In the present case, the continuous simultaneous dissolution and reprecipitation may be the most probable mechanism for this conversion The breakdown of the DCPD thermodynamic equilibrium at the DCPD- solution interface would take place due to the high undersaturation (σ = -0.330) Since monetite is thermodynamically more stable than DCPD at pH values around 5.0, the equilibrium phase would be monetite (Figure 5-9) 54 At the DCPD-solution interface, continuous dissolution of DCPD occurs, whereby the enrichment of Ca 2+ and HPO4 2- ions occur, and thermodynamic equilibrium is reached and monetite is precipitated This is in agreement with the previous results, and suggests that monetite crystals may appear during the formation of the callus 55

Samples taken after 300 min of the DCC experiment (Figure 5-8c) showed calcium oxalate monohydrate with traces amounts of monetite and calcium oxalate trihydrate (COT) Nucleation and growth of calcium oxalate indicated that COT was the hydromorph most likely to form as an initial crystal nucleus (Eq.5-10); COT is thermodynamically less stable than COM In accordance with the Otswald-Lussac Law, which states that the most soluble form will precipitate first then follow sequentially to the least soluble Previous research 56 showed that the dissolution rates of calcium oxalate n-hydrates differ markedly, with second-order rate constants at relatively high under-saturation falling in the order k(COT) > k(COM) > k(COT) Transformation of the higher hydrates to COM is considerably more rapid for COT, and the evidence indicates that surface controlled crystallization and a relatively low rate of COD growth results from the fact that this phase does not precipitate directly from pure calcium oxalate

Figure 5-9 54 Solubility isotherms, pCa vs pH of solutions saturated with respect to various calcium phosphate, Ca(OH)2-H3PO4-H2O at 25 o C solutions Thus, initiation of COT nucleation followed by a period of crystal agglomeration, proceeds until the solid state transformation takes place (Eq 5-10) Crystals of calcium oxalate monohydrate are then in contact with a slightly supersaturated solution of calcium oxalate and consequently undergo crystal growth according to Eq 5-11

The small number of studies using urine that report the presence of COT may be due partly to its relative stability under certain conditions, such that in experiments with long incubation times COT transforms to COM as described in studies involving inorganic growth media [57-60]

Calcium oxalate dihydrate (COD) is not a factor in the present investigation and no evidence has been found to indicate that the dihydrate form can result from the transformation of unstable higher hydrates [61]

Discussion 173

The present studies clearly demonstrate that brushite crystals may serve as an effective substrate for the heterogeneous nucleation of COM at supersaturations that normally do not lead to spontaneous precipitation During brushite dissolution, monetite has also been observed by X-ray diffraction This further demonstrates that the phase transformations of DCPD monetite and COT COM can occur under the previously mentioned in vitro conditions

Tang and Nancollas 62 also pointed out that brushite surfaces are good substrata for COM formation, owing to high local calcium concentrations It is then possible for these COM crystallites to aggregate to form concretions that could ultimately evolve into kidney stones The present in vivo experiments suggest that the brushite is located at the core of COM crystals, a position frequently encountered in kidney stones The initial site of formation of Randall’s plaques was recently reported to be in the basement membranes of thin loops of Henle This finding is of particular importance, as this is the earliest site within the nephron where high calcium phosphate supersaturation is encountered, and under slightly acidic conditions would favor brushite formation Furthermore, collagen, such as that in basement membranes, has been shown to precipitate brushite from urine

As brushite is thermodynamically less stable than hydroxyapatite, conversion to hydroxyapatite will be expected over time The hypothesis that Randall’s plaques are important in the initiation of CaOx stone disease is consistent with the observation that the number of CaOx stones formed is consistent with to the amount of papillary surface covered by plaque 63

This hypothesis is also consistent with our previous observations showing that COM is nucleated by apatite These calcium phosphate crystals located within the kidney provide not only a site for the nucleation of CaOx crystals but also an attachment that allows them to increase to a size that will be retained within the urinary system This study also shows that in nature, biomineralization of a thermodynamically stable phase can be induced by another more readily precipitated inorganic phase, which can also control the characteristics of crystallization such as aggregation etc, see Figure 5-10

Figure 5-10 Schematic showing the dynamic process based on preferential association of calcium ions with oxalate to form COM and the release of phosphate from brushite performed at lower levels of supersaturation than those in urine.

Conclusion 176

Brushite crystals serve as a very effective substrate for the heterogeneous nucleation of COM at a level of COM supersaturation at which spontaneous precipitate of this phase does not occur During brushite dissolution, monetite was confirmed by x-ray diffraction The present DCC data obtained under well controlled experimental conditions provide a physicochemical explanation for the initiation of COM crystallization, and also for the relative paucity of calcium phosphate detected in the majority of CaOx renal stones As protein inhibitors of crystallization within urine were not present, our studies could be performed at lower levels of supersaturation than those present in vivo It should be noted that results of studies of the effects of protein inhibitors on the nucleation (and growth) of COM on calcium phosphate substrates will be of obvious interest.