Colloid and Crystal Formation in Parotid Saliva of Cystic Fibrosis Patients and Non-Cystic Fibrosis Subjects. I. Physicochemistry

578 ALLARS ET AL

15. Mandel, I. D . Kutscher. A.. Denn~ng.C. R.. Thompson. R. H.. and Zegarell~,E. 512 (1964). V.. Saliv:~rystudleb in cystic fibrosls Amer. J. DIS. Child . 113: 431 (1967). 23. Wallach, D., and Schramm. M.: Calcium and the exportable proteln In rat 16. Mandel. I. D, Thompson. R. H.. Wotman, S. Taubman. M., Ku~scher,A. H.. parotid gland: Parallel subcellular distribution and concomitant secretion. Zegarelli. E V.. Denn~n?.C. R . Botwick. J. T . and Fahn. B. S.: Parot~dsaltva Eur. J. Biochem..2/-433 (1971). in cystlc librosls. 11. Electrolytes and prote~n-boundcarbohydrates. Amer. J. 24. Wiesmann, U. N.. Boat, T. F., and di Sant'Agnese. P. A.: Sodium concentration Dis. Child.. I10 646 (1965). in unstimulated parotld sallva and on oral mucosa in normal subjects and in 17. Mangos. J. A.. McSherr). N. R.. and Benke, P J.: A sodium transport inh~bitory pattents with cystic fihrosis. J. Pediat., 76: 444 (1970). lactor in the saliva of pattents with cystlc f~hros~sof the p:increas. Pediat. Res.. 25. Wotman. S., Mandel, I. D.. Mercadante, J., and Denn~ng,C. R.: Parotid and 1. 436 (1967). submaxillary calcium In human cystic libros~s.Arch. Oral B~ol. 16 663 (1971). 18 Marmar, J., Bdrbero. G. J . and Sib~nga.M S. 'The pattern of parotid gland 26. Wotman. S., Mercandante, J., Mandel. I. D., Goldman. R. S., and Dennln?. C : secrctlon in cystic fibrosis of the pancreas. G:~stroenterolog). 50: 551 (1966). The occurrence of calculus in normal children, children w~thcystic fihrosis. and 19 ,Mattheus. E. K.. Petersen. 0 H.. and W~ll~ams.J. A,: Pnncrcaltc acinar cells: children with asthma. J. Pcriodontol.. 44: 278 (1'173). Acetylchol~ne-~nducedmemhr:~ne depolnr~zat~on,calc~um efflux 2nd myla lase 27. H & l Instrumentation, Teaneck, N. J. rele~seJ Phys1o1. 234: 689 (1973). 28. Phadehas Amylase Test. Pharmac~aAB, Uppsala. Sweden 20. Saggcrs, B. A,. Lawson. D.. Stern. .I., and tdgson, A. C.: Raptd method for the 29. Computer Design Corporation, Los Angeles. Callf. detection of cystlc fibroa~aof thc pancreas in ch~ldrcn.Arch. Dis. Childhood. 30. Thls research was supported in part by The Clive and Vera Ramac~ott~ 42. 187 (1967). Foundations and The Cystic Fibros~sAssociation of New South Walcs. 21 Schneyer. L. H.. Young. J. A., and Schncyer, C. A,: Salivar) secretion of 31. Requests for reprints should be addressed to: J. Blomfield. M.Sc., Children's electrolytes. Physiol. Rev.. 52. 720 (1972). Med~calResearch Foundation, P.O. Box 61. Camperdown. N S.W. 2050 22. Shackleford, J. M.. and Bentle!, H. P. Carbohydrate histochem~atry of the (Australia). salivar) glands and pancreas in cyst~clibroa~s. J. Histochem. Cytochem.. 12: 32. Accepted for publication December 18. 1975.

Pediat. Res. 10: 578-584 (1976) Calcium parotid saliva cystic fibrosis protein inhibitors turbidity inorganic phosphate Colloid and Crystal Formation in Parotid Saliva of Cystic Fibrosis Patients and Non-Cystic Fibrosis Subjects. I. Physicochemistry

HELEN M. ALLARS, JEANETTE BLOMFIELD,'3" ANNE K. RUSH, AND JOHN M. BROWN Children's Medical Research Foundarioti atzd Cystic Fibrosis Clinic, Royal Alexandra Hospiral for Children, Svdne.~,New Sourh Wales. Australia

Extract Parotid saliva from patients with CF and control subject\ ih clear as collected, but saliva with elevated protein (aniylase) and Two types of turbidity were found in parotid saliva from both on standin? (4,, c\.~,,~ cystic fihrosis C'F) patients and non-C'F hubjects. 011 cooling salita. fibrosis patients tend to have parotid saliva with higher am\.lase a rapidly forming, rerersible, cold-dependent turbidity appeared in and concentrationsthan age.matct,ed control huh. increabing amounts with decreasing temperaiilrc aiid iiicreasiiig jects (4, 6), and turbidity formation is therefore encountered more protein concentration. At 37", a slowly forming, stahle turbidity frequently and to a greater degree in parotid saliva of CF patients. appeared in increased amounts in parotid saliva samples containing Aspects of physicochemical properties of this turbidity formatron. increased amounts of calcium. The 2" centrifuged pellet consisted particularly in relation to temperature and to protein, calcruni. and predominantly of protein, whereas the 37" pellet contained calcium, phosphate concentrations,are recorded in paper, and electron inorganic phosphate. and protein. The cold-dependent turbidity at micioscopc L;ppzarancc and clcctiophcjre:ic propertie!, :he 2" was not inhibited hy EDTA, but 37' turbidity was dramatically insoluble material are reported in a subsequent paper (2,. ln inhibited. Urea and guanidine hydrochloride reduced 2" turbidity, paper only the nature of the turbidity has been not its and, to a lesser extent, inhibited 37" turhidity. I'he tendency towards relativeoccurrence in each group of subjects, higher levels of protein, amylase, and calcium in C'F compared with child control parotid salira (4, 6) causes a greater incidence and degree of turbidity formation in salira of C'F patients. In this paper SUBJECTS AND METHODS only the nature of the turhidity has been investigated, not its relative occurrence in each group of subjects. SAL.IVA COI L.ECTION Subjects for the study were 13 cystic fibrosis patients aged 8 13 Speculation years (attending the Cystic Fibrosis Clinic at the Royal Alexandra In cystic fibrosis, a tendency towards high levels of calcium and Hospital for Children), and 7 young adult non-CF subjects aged protein in parotid saliva would mediate towards deposition of 18-24 years (hospital staff). To some extent subjects were selected colloidal protein and calcium phosphate within parotid gland ducts from those known to have a high parotid saliva flow rate and high and on tooth surfaces adjacent to duct orifices. A similar overse- protein and calcium concentrations, in whom the saliva had been cretion of calcium and protein in other C'F exocrine secretions may noted in previous studies (4, 6, 7) to become turbid on standing at result in obstruction by stagnant or cooled secretions. room temperature. COLLOID AND CRYSTALS IN CF SALIVA. I 5 79

Saliva was collected using a modified Lashley cup attached by ing increased amounts of calcium. Aspects of phqsicochemical suction over the orifice of Stenson's duct. Stimulation was 2 drops behavior of turbidity formation on cooling below 37" (AOD:,,:7) of 5% citric acid on the tongue every 30 sec for periods up to 10 and with time at 37" (AOD:,::) were investigated. min, which gave sufficient volume for tests (at least 3 ml). In- formed consent was obtained from adult s.uhjects and from pa- rents of children in this study. The cold-dependent turbidity formed immed~atelyon cooling parotid saliva to 2", and disappeared instantly on rewarming to 37" (Fig. I), demonstrating the reversible nature of the complex The turbidity of parotid saliva was measured as the optical causing turbidity. densit) of a sample in a quartz semi-microcuvette (path length I The turbidity increased rapidly with decreasing temperature. cm) at 320 nm in a Gilford 240 spectrophotometer. Cuvettes forming two types of temperature-dependence curves (Fig. 2): one incubated at temperatures below room temperature were coated being exponential (A)and the other having a central plateau reglon with Calotherm liquid demister (nonultraviolet absorbing) before (B). The graph of log AoD&,:' v.5 I/T(OK I) x lo4 (Fig. 3) was reading to prevent water condensation on the outside of the cold linear for A and consisted of two intersecting linear regions cuvette. ("bent") in B. Cold-dependent changes in the turbidity of the saliva were There was a direct relationship between the temperature-de- measured after 10 min of temperature equilibration of aliquots of pendent turbidity change and the number of particles causing light saliva in cuvettes in water or ice water baths at temperatures of scattering and hence also for the equilibrium constant (K) of col- 37". 25". 20". ISo, 5", and 2". Results were expressed as the loid formation. Using- the Van't Hoff thermodvnamic relationship optical densit! difference at 320 nm between the reading at dlogK AH temperature (T)and the initial optical density at 37' (AODj;,37).To (== =) the positive linear slope of the plot (AOD;.,"' give a thermodynamic relationship, the optical density difference vs I/T("K '))was proportional to the enthalpy change (AH) for was plotted against the inverse of temperature (1/T x 10' "K '). turbidity formation, i.e.. colloid formation involved loss of energy Reverhibility of cold-dependent turbidity was assessed by chang- (AH < 0). Since the system was at equilibrium (AG = 0. AG = ing saliva s;rniples at 10-min intervals from 37" to ZU and back to AH - TAS). then the entropy change (AS) as colloid formed was 37" several times, and comparing OD,,,, with the readings of negative. indicating production of a more ordered systern. The aliquots maintained at 37" and 2". linear "bent" lines may have been due to particles coalescing The turbidity at 37" was determined by measuring OD,,,, of rather than the formation of new aggregated particles. aliquots of parotid saliva incubated in cuvettes at 37' for various The amount of turbidity formed at lower temperature3 (10". 5". time periods up to 2 hr. and results were expressed as the optical and 2") was proportionally related to total protein concentration in density difference between the reading at time (t) and the initial the parotid saliva (Table I). There was no relationship with total optical density at rero time (AOD:;;:). calcium or Inorganic phosphate

4I)I)ITION 01. INHIHI IORS 01' TURBlDlrY

Paired aliquots of parotid saliva (0.9 ml) were made up to 1.0 In contrast to the cold-dependent turbidity. at 37' a slouer ml with (I) wirter as control; (2) solutions of EDTA (pH 7.0) to process of turbidity formation occurred with time and continued to give final concentrations of 0.5, I, 2, or 3 mM EDTA; or (3) solid rise for 2 hours or more (Fig. 4). At lower temperatures, an increase urea or gu~inidinehydrochloride to give final concentrations of I or in turbidity with time was slower, although cooling also caused an 6 M. immediate increase in OD,,,, because of the forniatlon of cold- Thc optical density at 320 nm (OD,,,) was read at 37O, then one dependent turbidity. At 37" the amount of turhidit! produced at of the pair of duplicates was incubated at 37O and the other at 2O different time intervals increased with increasing parot~dsaliva for 2 hr, and AOD:;:; and AOD:',,:7 were measured. total calcium concentrations (Table 2). There was no s~gnificant correlation between turhidity and tot~11inorganic phosphate or CIitMI('A1 ASSAYS protein concentrations. For chemical assays of turbid material, aliquots of saliva were centrifuged at the incubation temperature at 4,500 rpm for 10 min. The superniitant was removed and the centrifuged pellet was dihsolved in 4 M hydrochloric acid to half the original volume. Both pellets and uncentrifuged parotid saliva (for total concentrations) mere assayed for protein, calciunl, and inorganic phosphate content. Protcin wiis assaycd by :hc methud ::I' Lowry u! al. (2!). with tartrate being replaced by citrate (15). and pellet samples wcre neutralized with 4 M sodium hydroxide before assay. Calcium was determined by atomic absorption spectrophotometry (16) with samples diluted 1:10 in lanthanum chloride (1.2 5.000 ppm). Inorganic phosphate was measured using acid-ammonium molyb- date-ascorbic acid reagent ( 12). Pellet results wcre expressed as concentration in the original saliva volume.

K ESU LTS Two types of turhidity were found in parotid saliva from both ,20 40 60 80 100 120 CF patients and non-CF subjects, with no apparent qualitative TIME (MINUTES) differences between CF and non-CF saliva. On cooling saliva. a Fig. I. Reversibility of cold-dependent turhidity formation in parotld rapidly forming, reversible, cold-dependent turbidity appeared in saliva of a cystic fibrosis patient (calcium 0.87 mM; protein 365 rng/100 increasing amounts with decreasing temperature and increasing ml). Turbidity (O1>,,,) is plotted against time. Saliva kept at 37' (A): protein concentration. At 37", a slowly forming. stable turbidity saliva kept at 2" (e).Sal~va changed from 37" (A) to 2" (0)several tlmes appeared in increased amounts in parotid saliva samples contain- showing formation and clearing of turhidity. 580 ALLARS ET AL

TEMPERAT'JRE 1°cl Fig. 2. Temperature dependence of cold turbidity in parotid saliva. Turbidity, expressed as the change in OD,,,, at temperature (T)compared with reading at 37" (AOD;?,3'), is plotted against temperature (T). A. exponential curves: B, curves w~thcentral plateau.

Cystic fibro- Fig. 4. Turbidity formation in parotid saliva at 37'. Turbidity, ex- sis pressed as the change in OD,,, at time (t) compared w~threadlng at zero (CF) or Calcium, Protein. time (AOD;:), is plotted against time (t). Figure Symbols Non-CF mM mg/l00 ml Cystic fibros~s 2A 0---0 C F 0.75 230 (CF) or Calc~urn. Protein. H... .. Non-CF 1.03 255 Syrnbols non-Ct' rnM mg/ I00 ml A A Non-CF 0.85 165

Table 2. Dependence qf 37" turbidity or1 calci~tn~cor~cetltratiotr

Time at 37", X Y min No. r P

- Total AOD:,,;" 15 20 0.37 NS' calcium 30 22 0.61 <0.001 45 20 0.58 <0.01 60 22 0.72 <0.001 90 10 0.65 <0.05

r I20 6 0.77 NS - - % OK lo4 I Not significant. Fig. 3. The Van't Hoff relationship for cold-dependent turbidity for-

mation plotted as log AOD:;:" against inverse of temperature (1 /T OK' x 10'). A, from exponential curves of Figure 2A; B, from curves with central plateau of Figure 2B. COivlPOSiTiON 01. Pi~.l.l.tTSFROb1 I',IKOTID SALIV:, 21 ':'I I< IN('UHATlON AT 2" AN11 17" After incubation at 2°, the centrifuged pellet consisted predolni.. Table I. Depc.ndence of cold turbidit, on proteir~c.ot~cetrtratiotl nantlv of rotei in. with onlv sm~lluuantities of calcium and inorganic phosphate (Table 3). In contrast, in the 37" pellet there Temperature, were statistically significantly (P i 0.001) increased levels of X Y " C No. r calcium and inorganic phosphate, with approximately the same level of protein (Table 3). The Ca/P ratio ofthe 37" pellet was also Total AOD"320 37 25 18 0,26 NS, 20 0,19 NS significantly (P 0.001) greater than that of the 2" pellet. protein 20 < 15 20 0.34 NS EFFECT OF TUI'AI. CALCIUM CC)N('I:NTKATION ON PI-1 I.I:1 10 19 0.46 <0.05 COMPOSITION 5 20 0.50 <0.05 2 19 0.65

Table 3. C'onlpositiotl of pel1er.c /rot71 parotid sali~'aoffer itlcuhatiotr at 2" fbr 10 t~litrntrd ar 37" fhr 2 hr'

Protein, l norganic Temperature, "C No mg/ I00 ml Calciurn. mM phosphate, mM Ca/P ratio

' Pellet concentrations were calculated per volume of saliva. and expressed as mean + SD. NS: not s~gnificant

Table 4. Relarionships hetweerr total proieirl corrcet~trariotiarrd pellet cor~iposiiionat 2" atld 37"

2" 37"

X Y No. r P No. r P

Total Pellet 23 0.37 NS' 22 0.33 NS protein calcium Pellet 23 0.07 NS 22 0.34 NS phosphate Pellet 23 0.78 <0.001 22 0.65 <0.01 protein

I Not significant.

Fig. 5. The relationship between parotid saliva total calcium concentration and centrifuged pellet calcium, inorganic phosphate, and protein content after 10 min at 2' and afkr 2 hr at 37". Cystic fibrosis and non-cystic fibrosis subjects.

37" pellet (Fig. 5). The regression lines at 2" were appreciably lower than those at 37' (Fig. 5).

EFFtCT Ot- TOTAL PROTEIN AND PHOSPHATF: CONCENTRATIONS ON PELLET COMPOSITION Parot~dsal~va total proteln correlated pos~t~velyw~th pellet protein concentration at 2O (P < 0.001) and 37" (P < 0.01) (Table 4). Total protein did not correlate with pellet calcium or phos- 055 0'5 phate. PELLET CALCIUM(~M) Parotid saliva total inorganic phosphate concentration bore no Fig. 6. Interrelationships between parotid saliva centrifuged pellet con- significant relationship to pellet components at either temperature. tents of calcium, inorganic phosphate, and protein after 10 min at 2" and after 2 hr at 37'. Cystlc fibrosis and non-cystic fibrosis subjects. INTERRELATIONSHIPS OF PEI.1.ET COMPONENTS

At 2", inorganic phosphate and protein deposition in the pellet EFFECT OF EDTA ON PEI.1.ET COMPOSITION were uositivelv related (P < 0.05). and neither of these comoo- nents has relaied to calcium in the bellet (Fig. 6). After incubation The production of cold-dependent turbidity at 2" was not at 37", however, calcium in the pellet was positively correlated inhibited by the addition of EDTA (0.5 3 mM). The protein with both pellet protein (P < 0.001) and inorganic phosphate (P < concentration deposited at 2" remained approximately the same. 0.001), as was pellet protein with pellet inorganic phosphate (P < whereas calcium and inorganic phosphate in the pellet remained 0.001) (Fig. 6). low or decreased. In contrast. 37" turhidit? was dramat~callq inhihiteel b! the colloidal aggregates of parot~d protcins formed in incrc;~\cd presence of EDTA in the same concentration range. W~thEDTA amounts in saliva with elevated protein concentration\. and add~tionthe d~ffercncein OD,,,, al'tcr 2 hr remained at /em. in independently of the avi~ilithilityof ci~lcium. contrast to r~singlevels In the controls. Thc 37" pellet calcium and The Van't Hoff relationship for turh~d~t!forrn:~tion indicated inorganic phosphate levels were lowered to level\ claw to /cro, and that energy was lost (AH .- 0) dur~ngprotein aggrcgatlori. \rh~le pellet protein conccntration in most caws \fa\ lowcrccl signifi- the negative entropy change (AS c, 0) suggested that thc process cantl? in the prewnce of t.DT?\. produced a more ordered arrangement 01' proteins. There \+a\ no correlation hetween slopes of lines obtaineel in the plot 01' log AOL)':;:,'v.r I/T("K ') and total calcium. protein, and pho\phatc concentration or thc group of subjects uwd. i.c.. thc energ! Both urea (I and 6 M)and guanidine h!drochloriclc ( I and 6 M) requirement5 (AH) for turbidit! forniat~on \rere not rel;~ted reduced cold-dependent turbidit! at 2" b> louering thc pellet directly to any one of these factors. In cascs where the line:tr protein. calcium. and inorganic phosphate concentrations. The relationship was "bent." there may be a nonequilibrium situation effect \+as greater for guanidine hydrochloride than I'or urca. with particles coalescing in prefere~ceto forrnat~onof neb\ parti- Urea and guanldlne h>drochloridc at 6 bl conccntr:itions also cles, and the consequent effect on turbidity could change the tem- inhibited 37" turbidity. whereas there was less inhibition at I M perature dependence ofthe colloid's equilibrium constant. concentration\. Inhibition \\:is evident as decreased change in (iuitnidine hydrochloride (I and h M) wirs an el'l'ccti\e ~nh~hitor OD,,,, and diminished pellet content after 2 hr incubation at 37" of cold-dependent turhidity, effectively maintaining clear parotid when compared uith controls. saliva at 2". Comparable concentrations of urea \vcrc not :is effective in inhibiting colloid formation. although urca die1 produce s~gnificnnt reductions in turbidity. (iuanidine h!drochloriclc is thought to break down hydrophobic interactions :~nd hence denature proteins and separate subunits. \rtiercas urca h;i\ been These studlcs of citric acid-stimulated parotid \aliva have found to enhance hydrophobic bonding (24), the re;i\on I'or it\ disclosed formation of t~odifferent t!pes of turhidit!. one forming preferentially at ZU and the other at 37". The ?" turbidity was den~rturi~tingeffect being unknoun. The h!drophoh~c bonding ol' the colloid in parotid saliva is prohahl! dc\tro!eti b! guan~dinc rapidl! forming. uas reversible on reuarming. contained colloidal hydrochloride. resulting in disaggrcgitt~on the protein\. \.. hercas proteln, and correlated poitivcl! with parotld sali\a protein 01' urea, because of its polar nature. may on1 destro! thc ion~cor concentration. The 37" turhldity formed slowly, was stable. contained a calcium-phosphate-protein complex. and correlated hydrophilic bonding of proteins and phosphate Ion\ cau\lng le\\ positivel! uith parotid saliva c:ilcium concentration. drastic destabilication of the aggregate. Since pcllct proteln and pellet phosphate concentrations correlated. phosphate ions may Although both tqpes of turbidity ma! occur In saliva of both C'F patients and non-CF subjects, the tendenc? towards higher levels bind specifically to colloidal aggregate\. cither \tabili/ing the surface of particles or forming electrostatic bonds hetween protein of protein, arnglase, and calcium in C'k. in comparison with child in the aggregates. control parotid saliva (4, 6) causes a greater Incidence and degree of turbidity formation in saliva of CF patients. The colloidal aggregates may have a partial ordered arrangement 01' proteins (AS < 0) possibly because of (I) regions of hydrophobic bonding between nonpolar portions of protclns. (-7) COI I OID t-ORM:\TION IN P:\UOTII) SAL-IVZ phosphate binding at cationic sites on proteins. and (3)thc surface The reversible nature of the cold-dependent turhidit! formation orientation of hydrophilic groups of proteins. suggested a physical equ~libr~urnof aggregating saliva components The protein composition of parotid colloid has been p;~rtiallc demonstrating a temperature-sensitive bonding. This was further characterized (2). showing preferential involvement of \ome pa- confirmed hq the temperature-dependence of the equilibriun~ rotid proteins, in particular "proline-rich" proteins and a phospho- constant for the complex. The increase in cold turhidity formation protein. with increasing saliva protein concentration and its independence Descriptions of protein micelles or colloid in phls~ol+~cfluids of total calcium and phosphate concentrations indicated that the are rare. The best described protein niicclle system is the casein process might involve aggregation of parotid protein. The compo- micelle (22. 23. 31). although recently a cold-dependent turbidity sition of the 2" pellets supported this theory since the! conta~ned has been found in boar seminal plasma (10. 16) (Tahlc ') and in appreciable amounts of protein and only small amount5 of calcium human pancreatic juice (I ). and phosphate. The addition of EDTA in amounts sufficient to The white opalescence characteristic of milk can he attr~hutcdto chelate all divalent ions had no effect on reducing either turhidity light scattering from sm:~llspherical ni~cellesof casein (0.04 0.30 or pellet proteln concentration. Electron microscopy of turbid pm diameter). Milk micelles do not form at O0 6". hut form material at 2" showed amorphous. round particle\ ranging in six preferentially at 37O and require calcium (31) (Table 5). from 0.1 to 20 pm (2, 4). Thus the cold turb~dityappeared to he Proteins in hoar seminal plasma form a tinc-stah~li/edcold-

Table 5. Comparison q/'ph~~sicochrt?~icalrequirenrrtrr.t /'or /brttrariotr o-f micelles it1 milk. zinc-pronrotrd opal~scrrlc~~it1 hour .ren~itrol pla.c.mu, rrtrd colloid /ormution in parotid saliva

Milk (22. 23, 71) Boar seminal plasmla (10. 26) Parotid sal~va

I. Micelles l'orm at 37": no formation I. Colloid opalescence forms at 4O; I. Colloidal turbidity increases on atO" 6" clears at 37O cooling to 2"; clears at 37O 2. Crit~calrninimum amount of 2. Zinc is essential for stability 2. Calcium is not essential for tur- calcium e\rential for stabil~ty ol'turbidity; t:I)TA and citrate bidity formation; LDTA does not clear turbidity cffect cold-dependent turhidity: guanidine hydrochloride and urea clear turbidity 3. Micelles contain calcium, ir,-. 0.. 3. Colloid consists of protein of 3. Colloidal aggregates consist of and K-caseinsbound in a stable molecular weight approx. 50.000 one or more parotid proteins and structure which binds zinc hound phosphate ions COLLOID AND CRYSTALS IN CF SALIVA. I 583 dependent turbidity (10, 26) (Table 5). The turbidity is present as incidence and severity of dental calculus on tooth surfaces ad-jacent an opalescence at 37" and increases on cooling, behaving in a re- to the ducts of both the submandibular-sublingual glands and the versible manner similar to parotid saliva, but being inhibited by parotid glands (32). chelation of zinc with EDTA. Turbidity in the cold increases with increasing zinc content and decreases with increasing EDTA or ci- RELEVANCE TO OBSTRUCTION BY tXOCRlNt SECRETIONS IN CF trate concentrations, but is independent of calcium. It has been suggested that the insoluble material of CF Thus, the colloid formation of parotid saliva is cold induced and submandibular saliva, i.e.. undissolved "spherules" (zqmogen independent of divalent cations, that of milk has reversed tempera- granule components) and hydroxyapatite crystals. could block the ture requirements and a calcium dependence, and that of seminal ducts of the subrnandibular gland (5). Similarly the material plasma is cold induced and zinc dependent (Table 5). formed at 37" in this study. i.e., hydroxyapatite crjstals and It is of interest that in CF seminal plasma higher concentrations associated protein, may obstruct the ducts of the parotid gland (30, of zinc. calcium, magnesium, and enzymes and decreased solubil- 28). Elevated concentrations of protein and calcium in CF ity of protein have been reported (27). Adult male CF patients are compared with control parotid saliva (4, 6) would indicate a usually sterile, probably because of atrophy of the vas deferens higher likelihood of duct blockage in CF patients. since turbidity subsequent to blockage by insoluble material (25). increases more rapidly at higher calciilm levels, as does the In postoperative human ductal pancreatic juice, two types of amount of insoluble complex formed. Unstimulated resting pa- protein precipitation have been observed (I). Protein concentra- rotid saliva has high calcium and protein levels (13, 14). It seems tion-dependent fine precipitates formed on chilling inactive pan- feasible that overnight in unstimulated glands of the CF child creatic juice and redissolved on rewarming to 37". In activated there may be deposition of the calcium-phosphate-protein com- pancreatic juice, flocculent precipitates were formed, which did not plex in stagnant saliva in parotid acini and ducts. In both CF and redissolve at 37". Cation dependence was not investigated. non-CF subjects, parotid calculi and dental calculus on teeth may Chronic ethanol administration to dogs has shown rises in have a similar origin. cholecystokinin-stimulated pancreatic juice protein concentration and output after 6 weeks, and Palls after 12 months (30). After 6 SUMMARY weeks, microcalculi (rich in protein and calcium) and protein plugs Two processes of turbidtty formation in parotid saliva have been were excreted from the main pancreatic duct, and it was considered described in this study. Colloidal aggregates of proteins formed at that these may be causing incipient pancreatic lesions because of temperatures below 37" in increasing amounts related to increasing an obstructive mechanism, thus leading after 12 months to reduced protein concentration and decreasing temperature. The formation pancreatic secretion. A similar type of obstructive mechanism may of colloid was reversible (clearing instantaneously at 37O), inde- be operative in the pancreas of the cystic fibrosis child. pendent of the presence of divalent ions, exothermic in nature, and resulted in a more ordered arrangement of proteins In the saliva. CRYSTAI FORMATION IN PAROTID SALIVA The colloidal particles (confirmed by electron microscopy (2)) consisted of several proteins (characterized by electrophoresis (2)). At 37". parotid saliva samples often showed a gradual increase and some specifically bound phosphate ions. There was relatively in turbidity, and analyses of centrifuged pellets indicated forma- little cold turbidity in the presence of urea or guanidine hydro- tion of insoluble calcium, phosphate, and protein. chloride. Analysis of 2-hr 37" turbidity with transmission electron The second type of turbidtty consisted of a calcium-phosphate- microscopy and electron diffraction has shown crystalline material protein complex which formed most rapidly at 37'. and more which was often identified as hydroxyapatite (2). However the slowly at lower temperatures. Turbidity, and consequently the pellet Ca/P ratio of 0.97 * 0.34 and the presence of amorphous amount of calcium, inorganic phosphate, and protein deposited material by electron microscopy suggests incomplete crystalliza- in the pellet after incubation at 37'. increased with increasing total tion at 2 hr. Change of pH because of loss of carbon dioxide calcium concentration. Protein in the pellet also increased propor- probably contributed to formation of insoluble calcium phosphates tionally to total protein concentration. The addition of EDTA to in saliva samples with sufficiently high Ca/P products to cause parotid saliva in amounts sufficient to chelate all divalent ions in- precipitation (29). hibited the formation of the 37" turbidity. The crystalline material Grcan (17, 18) found stimulated parotid saliva from normal contained hydroxyapatite (identified by electron n~icroscopy(2)) subjects to be saturated in respect to hydroxyapatite (Ca,,(PO,), and proteins (characterized by electrophorests (2)). (OH),), whitlockite (Ca,(P0,)2), and octacalciurn phosphate No qualitative differences were observed between turbidity (Ca,H(PO,),). The calculations were based on ionized calcium formation in parotid saliva of children with cystic fibrosis and the and ionized orthophosphate determinations. The tendency to- young adult non-CF subjects of this study. However, the positive ward higher calcium concentrations in parotid saliva of children correlations between cold-dependent turbidity and protein concen- with cystic fibrosis (4, 6) would induce a greater degree of tration, and between 37" turbidity and calcium concentratton, hydroxyapatite crystal and turbidity formation in their saliva. indicate a greater degree of turbidity formation in parotid saliva of The protein bound to the calcium-phosphate complex has heen CF children. This quantitati\;e difference is i-liiisisieiit with dn characterized (2) as a phosphoprotein (9) and as proline-rich pro- increased incidence and severity of parotid small duct obstruction teins (3). It was noted that the presence of enough EDTA to che- and dental calculus adjacent to Stenson's duct in children with late all the diavalent ions resulted in no precipitation of the cal- cystic fibrosis. cium-phosphate-protein complex at 37' (suggesting that protein was adsorbed to the mineral (19)), yet allowed the formation at 2" REFERtNCES AND NOTES of the colloid containing the same proteins. Urea and guanidine I. Allan. J., and Wh~te,T. 7.: An alternate mech:~nism for the forrrlatton of protein hydrochloride also diminished turbidity formation at 37O. plugs in chronic calcifying pancreatttt5. Digestton. 11: 428 (1974). The calcium-phosphate-protein complex formed in parotid 2. Allars, H. M.. Cockayne. L). I H., Blomfield, J.. Rush. A. R.. van Lcnnep. E. saliva with calcium concentrations exceeding 0.5 mM both in CF W., and Brown. J. M . Colloid and cry>t;ll formnttc>n in parotid ~allvaof cystic patients and non-CF subjects. Similar formation of hydroxyapatite fibrosis patients and non-cystic f~bro~~ssubject>. 11. Electron mtcroscopy and electrophoresis. Pediat. Rcs., 10- 584 (1976). was noted in CF subrnandibular saliva after standing at room 3. ken, E. A,. and Oppenheim, F. G: Genet~cpolymorphtsm of prollnc-rich temperature for 2 hr (5, 8) and in whole saliva of heavy calculus human salivary proteins. Science, 1x0: 1067 (1973). formers (I I) after 20 hr of incubation at 37O. Electrophoretically 4. Blomfield, J.. Allars, H. M., Rush, A. R., van Lcnnep. E. W..and Brown, J M: the turbid material of the CF submandibular saliva contained Parotid serous hypersecretion in cystic fibrosis. Aust. Paediat. J.. 10: 75 (1974). 5. Blomfield. J., Dasc:llu. J., van Lennep. E. W , and Brown, J. M.: Hypcrsecrct~on protein bands postulated to be derived from undissolved compo- of zymogen granules in the pathogenesis of cystic fibrosis. Gut, 14: 558 (1973) nents of zymogen granules (5). CF children have an increased 6. Blomfield. J.. Rush, A. R.. Allars. H. M.. and Brown. J. M.: Parotid gland -84 ALLARS ET .4L

function In ch~ldrenu~th qstic fihro\~\and ch~ldcorltrol subject\. I'cd1.11 Kes.. 20 Joh;~n\cn. P. G.. Anderson, ('. M.. and tl.~cl~rn.B: C>btic F~hro\i\ol the 10 574 (1976). pancre:ir. A general~\edd~sturh~~nce of uatcr and clcctrol!tc rnovcnlcnt In 7. Blomfleld. I.. Ku\h. :\ K . ~ndAllar,. 11. %I. Intcrrelatiun\h~p\hetuccn Ilou cyocrlnc tlssueh. I ancct, i- 455 ( 1968) rate. arn!ln\c. c.ilc~un~.\ud~urn. put.is\iurn. and Inorganic pho\pl~atc~n Inter\ ;II 21 Lowry. 0. H.. Rosehrough. N. J . Farr. A. I... ;~ndRandall. R .I.: Protc~n atudlt.5 of cltrlc ;icld \t~rnulatedparotid \.III\:I of !oun: adult\. \rch. Or.11 Blul. rne:lhuremcnt u~ththe Fol~nphenol reagent .I. HIOI ('hem.. IV3 265 (1051 ). (In presh ) 22 McKenlle. H. 4.: MII~pr,ltc~n\. .-\d\:~n. Protc~n Chcm.. 2.55 (1907). 8. Hlomfield. I.. ran I.cnnep. E. W . Shore). C D.. Mdlln. .\ S.. I).~\c;~lu.J.. and 77 Nohlc. R. W.. and M'nugh. I). F.: Ca\ein m~cclle\ I-ormat~onand \tructurc I Brown. J M llltr;~\tructurcof the In vltro formation of h)clrox!;ipat~te In Amer. Chem Soc. .I.. #Y7 2736 (1965). iuhni:~nd~hularaali\n of children aith c!\t~c f~hro\~\.Arch Or;~lHiol . lY 1153 24. 0:ikenfull. D. G. The cll'ccta uf urea and gu:~nid~neh)drochlor~de (in h!dro- (1974) phah~cinteraction\ Proc. Au~t.Blochem. Soc . 7 20 (1974). 0. Boat. T. t .. ii Ic\m.in. U N.. and P;illar~c~n~.J C: Puriflc.~tlon;~nd propertlei of 75 Oppcnheimer. E. H.. and tsterll. J. R. Ohaerrati~mson c)stic fihro\i\ 01' the the calc~um-prec~pit;ihIcproteln in \uhn~a\illnr! \~II\:Iof norrn:il ;ind c!>l~c pancreas. I' Devclopmcnt;il changes in the m.~lcgen~tal \!stem J Pcdi;~t.. 75- fihrosi? \uhjcct\. I'cdiat. Re> . ,Y 531 (1974) 806 (1969). 10. Boursnell, J. C.. and Roberts, T. K.: The role of rlnc in prom~t~ngthe opales- 16. Koherts. T. K.. and Bour\nell. J. C.: The 1mport;incc of rlnc in cool~nghoar cence and cold-precipit:ition of boar acmin,~l plasma J Reprod tert~l..36 semen J. Reprod. Fert~l..35- 595 (1973). 91 (1974). 27 Rule. A. H . Kop~to.L.. iind Shwachman. H.: Chem~calanalbi~i of ejaculatei I I. Cancro. L. P.. Monc\. t..Mnrt~n. D M.. ;ind Wcaber. I. M - The l'orm;~tionof from patienti u~thc)\tlc fihro51s Fert~lStcr~l.. 1. 515 (1970). hydrox)apat~tedur~ng incubation of saliva. 14DR Abstr.. p. 107 (1070). 28. Shackleford. J. M., and Bentle). H P.: Carhoh)drate hiatochemlatr! of the 17. Chen. P. S.. Tor~hara.T. Y.. and Warner. H. M~crodctcrm~n;itionof pho.;- sal~var)glands and piincreab In cyst~cf~hrtlr~\. .I. ti~\tochem.C)tochem.. I?: phoru\. :\n:il. Chem.. ?,Y 1756 (1956). 512 (1964). 13. Daues. C : The \ccretlon of rnagneiium and c.ilc~uni In hu~n;inp;irot~d sal~\n. 2'). Tcrrn~ne.J. D . and Ilnnei. E. D : Calc~umpho\ph:itc depoiition Irom hiilanced Carle5 Re\.. I 333 (1967) salt solutions. Calc Tlss. Rea.. 15. 81 (1974). 14. Dawec. C: The effect\ of flow rate .lnd duration of \tirnulatlon on the 30 Tiscornia. 0.. Palaiclano. G.. and Sarleb. H.: Effects of chron~c ethanol concentration\ of proteln and the main electrol!tcb In hurn;in parotid aallra. adrn~n~itrat~onon canlne e~ocrine pancreatic wcretlon: Further studlc\. Arch. Oral Biol.. 14. 277 (1969) Digestion. /I. 171 (1974). 15. Eggstein. M.. ,ind Krcut7. F. H Vcrgle~chcnde Untrrauchungcn rur quan- 31. Waugh. D F.. and Nohlc. R W.: Case~nm~ccllc\. Format~on and \tructure. 11. J. titativen Eiuei\\he\t~mmung im L~quorund Fiweiiiarmen L.ii\unsen. Klin. Amer. Chem. Soc.. 87: 2246 (1965). Wschr.. 33- 879 (1955) 32 Wotman. S . Merc;~dan~c..I.. Mandel. I D . (;oldman. K. S.. and 1)enning. C: 16 Go=. B S. An:ll!\l\ of metals In \al~\.ih! ,Itornlc :~hcorpt~onrpectrohcop!. I. The occurrence of calculu\ In norln:il children. ch~ldrcnw~th cbrtlc I'~bro\~\.and Calc~um.J Dent Res.. 44 885 (IL)hl children with asthma. J. Periodontol.. 44: 278 (1973). 17. Gr4n. P The htatc of calc~umand Inorg,inlc orthopho\phate In human sallbCi 33 Thi, rese:~rch uai supported In part h) The C!\tic F~hroiisA~\oc~;it~oii of Neu Arch. Oral Hiol.. /X 1365 (1973). South Wales and The Cl~veand Vera Karn;ic~ott~Foundat~ons. 18. Grdn. P.: Satur;~t~onof human saliv;~\r~th calcium phosph.~te\. .Arch. Or.11 B~ol, 34. Kcqueits for reprint5 \h~uldhe addre\xd to: .I. Hlomfield. M.Sc.. ( h~ldrcn'\ I8 1385 (1973). Medical Research Foundation. P.O. Box 61. Camperdoun. N.S W 7050 19. Hay. D. I.: The Intcractlon of human p;~rot~d\:~l~rar! protcln5 w~thhydroxyapa- (Auatr:~lia). tile. Arch Orill Blol.. 18- 1517 (1073) 35, Accepted Ibr puhllcatlon Decemher IX. 1975

Cop)r~ghtO 1976 Intern,~t~onalPediatric Rc\c;irch Foundation. Inc

Pediat. Res. 10: 584-594 (1976) Colloid electrophoresis crystals parotid saliva cystic fibrosis phosphoprotein

Colloid and Crystal Formation in Parotid Saliva of Cystic Fibrosis Patients and Non-Cystic Fibrosis Subjects. 11. Electron Microscopy and Electrophoresis

HELEN M. AI-LARS. DAVID .I. H. COCKAYNE, JEANETTE BLOMFIELD,"' ANNt K. RUSH. ERNEST W. VAN LENNtP. AND JOHN M. BROWN

Chrldrer~'.~Medical Kesearch Foundurrorr arid C).tric Fibrosi.~C'linic, Rojui Ale\utidi~~::ospiioljoi C'hildrer;, a::d Elecrrori Micro.~copeUnir and Deparinlent oJ Hi.~~ologyarid embryo log^.. University of Sj,drrej,,S~dnej,. New Sourh Wales, Australia

Extract dine hydrochloride, and EDTA, resulted in no crystals being observed. Selected area electron diffraction from individual crystals Centrifuged pellets of turbid parotid saliva from cystic fibrosis showed predominantly hexagonal, rectangular, or square diffrac- (CF)patients and non-U subjects, obtained from saliva kept at 2' tion patterns. The hexagonal and rectangular patterns could be for 10 min, hhd the electron microscope appearance of amorphous, indexed as coming from hydroxyapatite. A transition from the round particles, and were thought to be colloidal aggregates of or- hexagonal to the rectangular pattern and back to the hexagonal ganic material. Drops of turbid saliva, from samples incubated for pattern could be obtained from individual crystals tilted in the elec- 2 hr at 2" or 37", additionally contained discrete, electron-dense tron microscope. The square diffraction pattern may be from octa- crystals having well defined angular morphology: usually cubic, rec- calcium phosphate or brushite. tangular, or approximately hexagonal. The inhibitors, urea, guani- Polyacrylamide gel disc electrophoresis of the parotid saliva in-