S Y M P O S I U M R E P O R TAustralian Dental Journal 2008; 53: 281–285

doi: 10.1111/j.1834-7819.2008.00063.x

Remineralization of deep enamel dentine caries lesions

JM ten Cate*

*Department of Cariology Endodontology Pedodontology, Academic Center for Dentistry Amsterdam (ACTA), Amsterdam, The Netherlands.

ABSTRACT

Enamel remineralization is generally studied in superficial (up to 100 lm) lesions, but in vivo caries lesions may be tenfolddeeper. This article addresses the question whether deep lesions, and extending into dentine, can be remineralized underoptimal conditions and if this process is influenced by agents affecting calcium phosphate precipitation and dissolution.Lesions through enamel into dentine were first formed in thin sections and then continuously remineralized for periods up to200 days. With longitudinal assessment by transversal microradiography it was showed that remineralization throughoutthe depth of the lesion and into the dentine was possible, although this process is very slow. Fluoride and bisphosphonatetreatments affected mainly the deposition in the outer enamel. Although it was assumed that this would affect the diffusionof ions to deeper layers, the treatments had no impact on remineralization in the inner enamel or dentinal parts of thelesions. These findings are discussed with relevant theoretical considerations, and in their possible clinical implications.

Key words: Remineralization, dentine, minimal intervention dentistry, fluoride, bisphosphonate.

Abbreviations and acronyms: DEJ = dentino-enamel junction; GIC = glass ionomer cements; KHN = Knoop Hardness Numbers; MID =minimal invasive dentistry; QLF = quantitative light fluorescence.

(Accepted for publication 19 May 2008.)

INTRODUCTION

The past decades have led to major changes inrestorative and preventive dentistry. Various investiga-tors reported that early caries lesions can be remineral-ized from saliva. This process was studied in detail inlaboratory experiments and in clinical trials.1,2 Adhesivedentistry has resulted in new non-metallic filling mate-rials, such as composites based on acrylate resin andglass ionomer cements (GIC) based on an acid baseprecipitation mechanism. Restorations made withadhesive materials no longer require a large cavitypreparation, but merely the removal of the tissueaffected by caries. Moreover, the notion became gener-ally accepted that restorative intervention is generallythe beginning of a long sequence of re-restorations,often leading to crowns and implants, irrespective ofhow well the first filling was prepared.

The concept of minimal intervention or minimalinvasive dentistry (MID) has combined these threemajor (paradigm) shifts in operative dentistry andthis philosophy is currently accepted worldwide. MIDnow has its own materials, congresses and researchorganization.3

In caries prevention, most protocols are still builtaround fluoride, although improving oral hygiene, sugarsubstitutes and antimicrobials are also part of the more

comprehensive packages of non-invasive care. Newmethodologies of caries diagnosis have been developed,using the early changes of fluorescence induced by cariesin the tissue detected by Quantitative Light Fluorescence(QLF),4 or with chromophores produced by bacteria (inthe Diagnodent device)5 as indicators.

Remineralization of enamel and dentine is studiedfrom two perspectives. First of all, it is the process ofmineral deposition from saliva or plaque fluid filling upsmall enamel or dentine defects formed during thedemineralization episodes resulting from acid attack onthe tooth. The relative magnitudes of demineralizationand remineralization determine whether a tooth surfaceremains sound or a caries lesion develops. This lesionmay then increase in severity eventually resulting ina (deep) cavity. (For more details about this cariesequilibrium, see Featherstone.2) Alternatively, reminer-alization is studied and described as the repair ofestablished lesions. Such lesions have developed over along period but may be filled in with calcium phosphateswhen external conditions favour mineral deposition.This type of remineralization may either be completeand partial; when the mineral precipitating in the lesionis less soluble than the original tissue, this remineraliza-tion will help in preventing or limiting future tissue loss.

Remineralization of superficial enamel lesions iswell documented in hundreds of studies completed at

ª 2008 Australian Dental Association 281

numerous laboratories in the last century. Studieson the basic mechanism of remineralization and onmethods to stimulate this process have led to theconclusion that the caries preventive effect of fluoride isbeyond any doubt. This is partly attributed to theenhancing effect of fluoride on calcium phosphateprecipitation, hence remineralization.6

A topic that has received limited attention is whetherthere is a point of no return beyond which remineral-ization can or does no longer occur. Koulourides (oneof the pioneers in remineralization research), stated thatif caries has weakened the tooth structure to below ahardness of 150 Knoop Hardness Numbers (KHN),remineralization could no longer be achieved.7,8 Con-ceptually this was seen as the point where the mineralstructure is destroyed to the extent that reprecipitationof mineral on remaining hydroxyapatite crystallites wasno longer possible. In caries diagnosis using X-rays, theextent of caries is graded at various depth levels in theenamel and dentine, realizing that loss of tissue hasoccurred also considerably beyond the depth identifiedon the X-ray picture. The current consensus is thatcaries beyond the dentino-enamel junction (DEJ)should be treated with restorations, and lesions up tothat point should receive extra preventive care. How-ever, it was never studied whether deep lesions,extending into dentine, can be remineralized if suchlesions are subjected to a continuous remineralizationscheme. Obviously, such lesions should be protectedfrom mechanical damage. Ideally, a study with thisobjective should be carried out in vivo or using an insitu model. However, given the fact that remineraliza-tion is a very slow process, it seemed technicallyimpossible to complete such a study with volunteersubjects or with patients, within an acceptable timeperiod.

This article reiterates previous results from in vitroremineralization experiments of deep lesions, extendinginto dentine.9 The aim was to explore whether suchlesions can still be remineralized and how this couldbe affected by treatments that would stimulate or inhibitcalcium phosphate precipitation. These findings are thendiscussed from a theoretical and clinical perspective.

Experimental design and results

The experiment was performed in groups of 100 lmthick sections cut from ground bovine incisors. Sectionswere fully embedded in Araldite, and after setting of theresin the outer 200 lm of the enamel was cut with adiamond coated wire sectioning machine. This wasdone to remove surface enamel and to reduce theenamel thickness to the DEJ. Lesions through enameland into dentine were formed during 10–15 daysin individual solutions containing 1.5 mM CaCl2,0.9 mM KH2PO4 and 50 mM acetate buffer (pH 4.8),

with the addition of 0.1 ppm KF to prevent surfaceloss.10,11 Next the lesions were immersed in solutionssupersaturated and stochiometric to hydroxyxapatite(HAP), so as to achieve remineralization (1.5 mMCaCl2, 0.9 mM KH2PO4 20 mM HEPES buffer pH 7.0and 130 mM KCl). All steps in this process weremonitored and recorded by taking microradiographs, aswas the remineralization phase by taking microradio-graphs weekly during 200 days. Profiles were scannedat fixed positions of the specimens. Five independentspecimens were run at each condition.

Treatments tested were control (no additional treat-ments), weekly five-minute fluoride rinses with 1000ppmF, addition of 1 ppmF to the remineralizingsolution, and a five-minute single treatment with thecalcium phosphate precipitation inhibitor bisphospho-nate (as 2 mMethaneHydroxy-BisPhosphonate).

The primary findings of these experiments are givenin Fig 1. To average out variation between specimens inenamel thickness and overall lesion depth, data areexpressed as percentage repair in four zones: outerenamel, inner enamel, outer dentine, inner dentine. Therelative remineralization parameter shows that remin-eralization in dentine (panels C and D) may be up to 80per cent after 200 days, while remineralization in innerenamel (panel B) plateaus at around 40 per cent. Onlyin the outer enamel (panel A) is the remineralizationsignificantly affected by the various treatments, thusresulting in relative remineralization values between 40and 100 per cent. (For full experimental details andresults, see ten Cate.9)

Theoretical considerations

Whether remineralization occurs in enamel-dentinelesions extending beyond the DEJ depends on severalfactors. First, the concentration of mineral ions at thesite of precipitation should exceed supersaturation tohydroxyapatite. This requires that not all calcium andphosphate ions entering through the lesion pores havealready been precipitated in the layers closer to thesurface. Secondly, the site of precipitation requiresnuclei for precipitation, considering that substantiallylarger degrees of supersaturation are needed forde novo precipitation of apatite or precipitation ontoremaining organic matrix than on fragments of enamelor dentine apatite crystals. Detailed electron micro-scopic analysis of crystallites in various zones of thelesion confirmed that remineralization occurs bygrowth of existing crystals to dimensions larger thanthe original crystallites.12

Rate control

Considering the theory of crystallization kinetics,a precipitation at greater depth requires that the

282 ª 2008 Australian Dental Association

JM ten Cate

precipitation reaction is slow compared to mass transferof ions, thus allowing diffusion to supply ions throughoutthe lesion. If such a condition is met, the mineral ionconcentrations are uniform throughout the lesion.

Little data are available to confirm this assumption.With Arrhenius plots (temperature dependence ofreactions), we previously determined that the activationenergies of remineralization of subsurface lesions versussurface softened enamel were significantly different. Wethen concluded that diffusion processes were ratelimiting in the lesion remineralization (judging fromthe lower activation energies), while surface softenedlesion remineralization was controlled by surfacereactions.13 The current findings seem to contradictthese early observations.

Demineralization

More potentially relevant data are available on enameldemineralization, showing that its rate is determined bydiffusion processes, although differences in mechanismwere noted between demineralization in vitro andin vivo.14,15 Recently, we developed an artificial groovemodel, in thin sections, to simulate demineralization infissures. In this model we periodically monitored mineralloss with microradiography and pH and calcium

concentrations throughout the depth of the grooves withmicroelectrodes. The observed gradients in pH andcalcium activities pointed to diffusion inhibition fordemineralization even in the 250 lm wide grooves (tenCate and Buijs, unpublished data). Making this compar-ison it should be noted that precipitation rates areprobably 10 times slower than dissolution rates atrelevant super- and undersaturation conditions.

Enamel-dentine continuum

Even if enamel and dentine form a continuum within thetooth, their respective apatite crystallites differ in sizeand composition, reflected in differences in solubility. Inan aqueous environment this would eventually lead tothe complete dissolution of the dentine crystallites infavour of the enamel crystallites, a process known asOstwald ripening in crystal chemistry. Findings inaccordance with this principle were observed whenenamel and dentine (lesions) were de- and remineralized,respectively, when placed in juxtaposition.16

Matrix

For in vivo remineralization, additional mechanismscould play a role as indicated by in situ studies

120(a) (c)

(b) (d)

100

80

60

20

40

0

Rel

ativ

e re

min

eral

izat

ion

(%

)

Time (days)

100

80

60

20

40

00 50 150100 200 0 50 150100 200

Fig 1. Average relative remineralization, with time of remineralization, for the five experimental groups in the four zones of interest: a: outerenamel, b: inner enamel, c: outer dentine, and d: inner dentine (n = 5 per Group). Markers indicate the five experimental groups: – control (r),

– weekly 1000 ppm fluoride treatments (m), – continuous presence of 1 ppm fluoride (s), – single MHBP treatment (h), and – combinationtreatment: single MHDP treatment at start followed at 56 days by weekly 1000 ppm treatments (n). Relative remineralization is defined as the

percentage of mineral deposited at a given depth to fill in the mineral removed during lesion formation. (Reprinted with permission from JM tenCate; J Dent Res 2001;80:1407–1411.)

ª 2008 Australian Dental Association 283

Remineralization of deep caries lesions

completed by van Strijp and colleagues.17 In a compar-ison of various fluoride toothpastes they observed fullremineralization of dentinal lesions in many of theparticipating subjects, and small changes in the contra-laterally placed enamel lesions. This observation hintsthat remineralization conditions for dentine lesions maybe favourable compared to enamel lesions, although thefull explanation of this finding is yet lacking. One mayformulate the hypothesis that the demineralized organicmatrix of dentine may constitute a scaffold to enhanceremineralization.

Furthermore, non-collagenous components of thismatrix (SIBLINGs, osteocalcin, proteoglycans).18,19

may interact directly with crystal formation and crystalgrowth during dentine remineralization. In the contextof the current experiment, this all could suggest thatremineralization of dentine proceeds faster than forenamel and would create a concentration ‘sink’ beyondthe DEJ.

Modelling

Many of the abovementioned findings need furtherstudy. However, given the duration of remineraliza-tion studies of deep lesions, it seems worthwhile toconsider in silico experiments; computer simulationsof remineralization using model parameters that canbe determined individually or are available in theliterature (rates of apatite precipitation, dissolution,diffusion constants, etc.). Obviously, as with anysimulation model, this approach helps to identify orillustrate the relative importance of the individualsteps in a complex process, in this case diffusion,precipitation, tortuosity, etc. It is beyond the scope ofthis presentation to describe the numerical approach indetail or give all the equations for the separate steps inthe process (for details see ten Cate).20 Examples ofsuch computational exercises are included as illustra-tion in Fig 2.

Enhancing remineralization of deep lesions

Fluoride

Traditionally, the focus in the development of agentsaimed to enhance remineralization has been on fluoride.The presence of low levels of fluoride increases thedegree of supersaturation with respect to fluoridatedhydroxyapatite. This thermodynamic property is therationale for the enhancement of remineralizationby fluoride. This, however, could lead to excessiveremineralization of the surface layer of lesions andconsequently lesion arrestment.21 In the describedexperiments (Fig 1) the tested agents (fluoride andbisphosphonates) were found to give rise to divergingresults in the outer enamel, but these treatments did not

significantly affect precipitation of mineral in the innerenamel and dentine. Fluoride treatments, whether as1000 ppm topical treatments or continuously present at1 ppm, were both beneficial for repair of deep lesions,at least in the outer enamel.

Calcium

After analysing comprehensive literature data onin situ remineralization, saliva and plaque mineralion compositions and saliva flow dynamics, we haveproposed that, in addition to fluoride, calcium may berate limiting in remineralization.22 Since then many

(a)

(b)

(c)

(d)

dentineenamel

mineral

concentration (IP)

dentineenamel

mineral

concentration (IP)

dentineenamel

mineral

concentration (IP)

dentineenamel

mineral

concentration (IP)

Fig 2. Computer simulation of diffusion – precipitation processes inpore in enamel lesion extending into dentine. Two dimensionaldiffusion was modeled using Fick’s law of mass transfer, and

precipitation using the thermodynamic laws of dissolution andprecipitation (for details see ten Cate).20 Various conditions were

modeled: – a, b: fast precipitation resulting in concentration gradientin pore and preferential deposition in outer enamel – c: slow

precipitation resulting in homogeneous concentration and precipita-tion throughout the depth of the pore – d: preferential deposition ofmineral in dentine, resulting in concentration sink towards dentine.

284 ª 2008 Australian Dental Association

JM ten Cate

new products, including toothpastes and chewinggums, have been formulated with the aim to supplycalcium ions to the oral cavity.23,24 Recently, clinicalstudies and in situ trials have confirmed the potential ofthis remineralization approach.25,26 There seems scopeto also study such products with the aim to enhancedeep rather than superficial lesions.

Clinical aspects

Obviously, conditions in the mouth are very differentfrom the ideal remineralization conditions used in theexperiments described in this article. An importantin vivo challenge to the remineralization of deep lesionsare the periods of low pH in the plaque which worsenthe degree of mineralization rather than improving it.Moreover, the teeth are subject to mechanical forceswhich could break the surface of the lesion and create aretention site for bacteria. Once the surface is irrepa-rably damaged, clearly the chance for a non-invasiverepair of the lesion is lost.

CONCLUSIONS

The experiments described in this study show thatremineralization of lesions extending into the dentine ispossible. This process takes a considerably long time,which clinically is unacceptable. Nevertheless, studiesto further our knowledge in this field would contributeto the area of minimal intervention dentistry.

REFERENCES

1. ten Cate JM. In vitro studies on the effects of fluoride onde- and remineralization. J Dent Res 1990;69 Spec No:614–619.

2. Featherstone JD. The continuum of dental caries–evidence for adynamic disease process. J Dent Res 2004;83:C39–C42.

3. Mount GJ, Ngo H. Minimal intervention: a new concept foroperative dentistry. Quintessence Int 2000;31:527–533.

4. van der Veen MH, de Josselin de Jong E. Application of quanti-tative light-induced fluorescence for assessing early caries lesions.Monogr Oral Sci 2000;17:144–162.

5. Lussi A, Hibst R, Paulus R. DIAGNOdent: an optical method forcaries detection. J Dent Res 2004;83:C80–C83.

6. Marinho VC, Higgins JP, Logan S, Sheiham A. Topical fluoride(toothpastes, mouthrinses, gels or varnishes) for preventing dentalcaries in children and adolescents. Cochrane Database Syst Rev.2003;4:CD002782.

7. Koulourides T, Cueto H, Pigman W. Rehardening of softenedenamel surfaces of human teeth by solutions of calcium phos-phates. Nature 1961;189:226–227.

8. Koulourides T. Experimental changes of enamel mineral density.In: Harris RS, ed. Art and Science of Dental Caries Research.New York: Academic Press, 1968:355–379.

9. ten Cate JM. Remineralization of caries lesions extending intodentin. J Dent Res 2001;80:1407–1411.

10. ten Cate JM, Duijsters PP. Influence of fluoride in solutionon tooth demineralization. I. Chemical data. Caries Res 1983;17:193–199.

11. ten Cate JM, Exterkate RA, Buijs MJ. The relative efficacy offluoride toothpastes assessed with pH cycling. Caries Res2006;40:136–141.

12. Silverstone LM, Hicks MJ, Featherstone MJ. Dynamic factorsaffecting lesion initiation and progression in human dentalenamel. Part I. The dynamic nature of enamel caries. Quintes-sence Int 1988;19:683–711.

13. ten Cate JM, Arends J. Remineralization of artificial enamellesions in vitro: III. A study of the deposition mechanism. CariesRes 1980;14:351–358.

14. Featherstone JD, Duncan JF, Cutress TW. A mechanism fordental caries based on chemical processes and diffusion phe-nomena during in-vitro caries simulation on human tooth enamel.Arch Oral Biol 1979;24:101–112.

15. Arends J, Christoffersen J. The nature of early caries lesions inenamel. J Dent Res 1986;65:2–11.

16. Lynch RJ, ten Cate JM. The effect of adjacent dentine blockson the demineralisation and remineralisation of enamel in vitro.Caries Res 2006;40:38–42.

17. van Strijp AJP, Buijs MJ, ten Cate JM. In situ fluoride retention inenamel and dentine after the use of an amine fluoride dentifriceand amine fluoride ⁄ sodium fluoride mouthrinse. Caries Res1999;33:61–65.

18. Tartaix PH, Doulaverakis M, George A, et al. In vitro effects ofdentin matrix protein-1 on hydroxyapatite formation provideinsights into in vivo functions. J Biol Chem 2004;279:18115–18120.

19. Milan AM, Sugars RV, Embery G, Waddington RJ. Adsorptionand interactions of dentine phosphoprotein with hydroxyapatiteand collagen. Eur J Oral Sci 2006;114:223–231.

20. ten Cate JM. A model for lesion remineralization. In: Leach SA,Edgar M, eds. De– and remineralization of dental enamel.London: Inf Retr Inc, 1983.

21. ten Cate JM, Jongebloed WL, Arends J. Remineralization ofartificial enamel lesions in vitro. IV. Influence of fluorides anddiphosphonates on short- and long-term reimineralization. CariesRes 1981;15:60–69.

22. ten Cate JM. In situ models, physico-chemical aspects. Adv DentRes 1994;8:125–133.

23. Reynolds EC. Anticariogenic complexes of amorphous calciumphosphate stabilized by casein phosphopeptides: a review. SpecCare Dentist 1998;18:8–16.

24. Winston AE, Bhaskar SN. Caries prevention in the 21st century.J Am Dent Assoc 1998;129:1579–1587.

25. Iijima Y, Cai F, Shen P, Walker G, Reynolds C, Reynolds EC.Acid resistance of enamel subsurface lesions remineralized by asugar-free chewing gum containing casein phosphopeptide-amorphous calcium phosphate. Caries Res 2004;38:551–556.

26. Morgan MV, Adams GG, Bailey DL, Tsao CE, Fischman SL,Reynolds EC. The anticariogenic effect of sugar-free gumcontaining CPP-ACP nanocomplexes on approximal cariesdetermined using digital bitewing radiography. Caries Res2008;42:171–184.

Address for correspondence:Professor JM ten Cate

University of Amsterdam and Free UniversityAmsterdam

Academic Center for Dentistry Amsterdam (ACTA)Royal Netherlands Academy of Arts and Sciences

Louwesweg 11066EA Amsterdam

The NetherlandsEmail: j.t.cate@acta.nl

ª 2008 Australian Dental Association 285

Remineralization of deep caries lesions