The biomechanical behaviour of the hyalinized periodontal ligament in dogs during experimental orthodontic tooth movement

Orthodontic tooth movement depends largely on the biomechanical properties of the periodontal ligament (PDL). Force application on a tooth results in strain within the PDL and alveolar bone. This, along with subsequent  uid  ow, leads to cell strain and induces the expression of a variety of regulatory factors. In response, cell differentiation, remodelling of soft tissues, and local resorption and deposition of alveolar bone occurs. Together, these events ultimately lead to tooth movement ( Ten Cate and Nanci, 2003 ; Krishnan and Davidovitch, 2006 , 2009 ; Henneman et al. , 2008 ). This mechanism, however, is complicated by changes in the structure of the PDL during orthodontic tooth movement. It was hypothesized that these changes will alter the biomechanical behaviour of the PDL. After a  rst phase that represents the initial movement of the tooth within its socket, tooth movement generally shows a temporary standstill due to hyalinization of the PDL. This is caused by obliteration of the blood vessels and subsequent local necrosis. Immediate turnover of extracellular matrix and direct bone resorption are disabled until the hyalinized tissue has been removed by macrophages and replaced by new loose connective tissue. Only then are osteoclasts recruited to the area and direct bone resorption can start ( Pilon et al. , 1996 ; Nakamura et al. , 2000 ; Ten Cate and Nanci, 2003 ). The biomechanical behaviour of the hyalinized periodontal


Introduction
Orthodontic tooth movement depends largely on the biomechanical properties of the periodontal ligament (PDL).Force application on a tooth results in strain within the PDL and alveolar bone.This, along with subsequent  uid  ow, leads to cell strain and induces the expression of a variety of regulatory factors.In response, cell differentiation, remodelling of soft tissues, and local resorption and deposition of alveolar bone occurs.Together, these events ultimately lead to tooth movement ( Ten Cate and Nanci, 2003 ;Krishnan andDavidovitch, 2006 , 2009 ;Henneman et al. , 2008 ).This mechanism, however, is complicated by changes in the structure of the PDL during orthodontic tooth movement.It was hypothesized that these changes will alter the biomechanical behaviour of the PDL.
After a  rst phase that represents the initial movement of the tooth within its socket, tooth movement generally shows a temporary standstill due to hyalinization of the PDL.This is caused by obliteration of the blood vessels and subsequent local necrosis.Immediate turnover of extracellular matrix and direct bone resorption are disabled until the hyalinized tissue has been removed by macrophages and replaced by new loose connective tissue.Only then are osteoclasts recruited to the area and direct bone resorption can start ( Pilon et al. , 1996 ;Nakamura et al. , 2000 ;Ten Cate and Nanci, 2003 ).
The biomechanical behaviour of the hyalinized periodontal ligament in dogs during experimental orthodontic tooth movement S. H. Jónsdóttir , E. B. W. Giesen and J. C. Maltha Department of Orthodontics and Oral Biology, Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands Correspondence to: J. C. Maltha, Department of Orthodontics and Oral Biology, Radboud University Nijmegen Medical Centre, PO Box 9101, 6500 HB Nijmegen, The Netherlands.E-mail: j.maltha@dent.umcn.nlSUMMARY During orthodontic tooth movement, the mechanical behaviour of the extracellular matrix of the periodontal ligament (PDL) determines the cellular processes involved in turnover of the PDL and alveolar bone.This mechanical behaviour is the basis for fi nite element (FE) models and FE analyses.Five young adult male beagle dogs were used to test the null hypothesis that the mechanical behaviour of the PDL is identical in normal and hyalinized PDL.Therefore, tooth transposition was measured after standardized force application by super-elastic nickel titanium (NiTi) coil springs, exerting a constant force of 100 cN for 5 hours in both conditions.
A rapid transposition during the fi rst few seconds was found.However, it was signifi cantly less for hyalinized than for non-hyalinized PDL.Subsequently, a short-lived creep movement was found for hyalinized PDL, while creep persisted at the non-hyalinized sides ( analysis of variance and Tukey ' s multiple comparisons post hoc tests ).The results showed substantial biomechanical differences between hyalinized and non-hyalinized PDL at different time points (Mann -Whitney).This indicates that FE models in the study of long-term orthodontic tooth movement, which are based solely on the characteristics of normal PDL should be reconsidered.
If no hyalinization of the PDL has taken place, a rapid remodelling of the extracellular matrix is found in the compression areas ( Pilon et al. , 1996 ).The principal collagen type I  bres are lost within a few days and replaced by a loose connective tissue that contains mainly randomly orientated thin collagen  bres.
As cellular reactions during orthodontic tooth movement depend on the strain within the PDL ( Krishnan and Davidovitch, 2006 ), knowledge of the changes in the mechanical behaviour of the PDL caused by structural changes is essential for understanding this process ( Wakabayashi et al. , 2008 ).However, to date , only limited data are available on the in vivo behaviour of the normal PDL ( Limbert et al. , 2003 ;Kawarizadeh et al. , 2004 ;Jónsdóttir et al. , 2006 ).Finite element (FE) models used to analyse the effects of orthodontic forces on the PDL and alveolar bone are based on these normal physiological conditions ( Tanne and Sakuda, 1983 ;Yoshida et al. , 2001 ;Viecilli et al. , 2008 ) or on ex vivo measurements ( Pini et al. , 2002 ;Poppe et al. , 2002 ;Dorow et al. , 2003 ).No data on the mechanical behaviour of the PDL during the different phases in the course of orthodontic tooth movement are available .
The aim of the present study was therefore to test the null hypothesis that the mechanical behaviour of the hyalinized 2 of 5 and non-hyalinized PDL, and thus tooth transposition within the socket under standardized loading conditions is equal.The term tooth transposition is used to distinguish tooth movement within the socket without bone remodelling from the movement that involves bone resorption and deposition.

Preparation of the animals
Five young adult male beagle dogs (age d 1.0 -1.5 years) with a complete permanent dentition were used for the experiment.Ethical permission was obtained according to the guidelines of the Radboud University Nijmegen, The Netherlands.The animals were housed in groups of two or three animals per cage under normal laboratory conditions and were fed standard dog chow and drinking water ad libitum .
The mandibular third and fourth premolars on both sides were extracted, as well as the maxillary second and third premolars and the  rst molars, in order to avoid interference with the orthodontic appliance.After 3 months, custom-made titanium implants (height 10 mm, diameter 3.1 mm, and sandblasted) with locking screws were placed bilaterally in the mandible between the second premolar and  rst molar.The dogs were medicated with Albipen L A (ampicillin 100 mg/ml ; Intervet, Boxmeer, The Netherlands) after implantation.

Orthodontic appliance
Crowns for the second premolars and  rst molars were made 3 months after implant placement.A tube and a hook were soldered on the buccal side of each crown.The crowns were cemented in place with PanaviaEx dental adhesive (Kuraray Medical Inc., Okayama, Japan) and a suprastructure with a holder for a stainless steel sliding bar was  tted on the implant.Low friction silicone bearings with an internal diameter of 2.01 mm were  tted in the tubes on the crowns, and a 2 mm sliding bar was placed through the bearings and through the holder on the implant.The silicone bearings were then  xed into the tubes with Fissurit F (Voco, Cuxhaven, Germany).Finally , the holder for the sliding bar was  xed into the suprastructure with Ketac-Cem (3M Espe AG, Seefeld, Germany).This enabled bodily movement of the second premolar and  rst molar with virtually no resistance and an axial play of 0.01 mm.At day zero of the experiment, a super-elastic NiTi coil spring, exerting a constant force of 100 cN (GAC International, New York, USA) , was  xed to the hooks on the crowns to induce reciprocal movement of the second premolar and  rst molar.At day 6 , the coil springs were removed in order to allow the ligament to relax.

Tooth transposition measurements
On day 7 , tooth transposition was carried out on the second premolar.It was assumed, in most cases , that hyalinized tissue was present at that time ( Pilon et al. , 1996 ;Van Leeuwen et al. , 1999 ).This was veri ed by histology or time-displacement curves in the subsequent period.
The dogs were anaesthetized as previously.The sliding bars and the suprastructures were removed from the implants and a special suprastructure for the measurements was  xed on the implant ( Figure 1 ).On this new suprastructure , a displacement transducer with a centric hole of 2.0 mm in diameter (MSH 707 ; Sony Magnescale Inc., Saitama, Japan) was  xed with Clear l ™ AP-X (Kuraray Medical Inc.).A magnetic ruler (SR-721 SP ; Sony Magnescale Inc.) with a diameter of 2.0 mm was inserted through this centric hole and through the tubes on the crowns and  xed at the second premolar by a locking screw.The magnetic ruler could then move through the transducer with the movement of the second premolar.The magnetic ruler was divided alternately into north and south magnetic poles, 200 µ m apart, which changed the electromagnetic  elds within the transducer by displacement of the ruler.The transducer was connected to a digital counter (LH 20C ; Sony Magnescale Inc.) to calculate the movement of the magnetic ruler passing through the stable transducer and thus the distal displacement of the second premolar with an accuracy of 0.5 µ m.
Figure 1 The experimental appliance with the measuring system. 1. implant, 2. crowns, 3. elastic, 4. displacement transducer, 5. magnetic ruler, 6. digital counter, and 7. personal computer .Displacement of the second premolar was induced on both sides by pre-stretched elastics (Z-pak elastics, Ormco® Orange, California, USA) that were selected to exert a force of 100 cN.The transposition was sampled at 1 Hz for 5 hours.

Hyalinized or non-hyalinized PDL
Two dogs were sacri ced for histological evaluation during general anaesthesia by a lethal dose of narcovet (sodium pentobarbital 60 mg/ ml ; Apharmo, Arnhem, The Netherlands).The second mandibular premolars and their surrounding bone were dissected.After  xation in 4 per cent buffered formaldehyde solution in 0.1 mol/l phosphate buffered saline and decalci cation in 20 per cent formic acid and 5 per cent sodium citrate, the tissue blocks were dehydrated and embedded in paraplast (Monoject Scienti c, Athy, Ireland).Serial mesiodistal section s of 7 µ m were cut and stained with haematoxylin and eosin.If hyalinized tissue was present in the PDL, the tooth was considered to be in the hyalinization phase.
The three remaining dogs were prepared for further orthodontic tooth movement, and time-displacement curves were constructed, based on intra-oral measurements.The teeth were considered to be in the hyalinization phase if they showed no movement for at least 2 weeks.

Data analysis
Time-transposition curves were constructed from the 5 hour measurements for the second premolars at both sides of each dog.Two different categories of curves were found; one showing limited tooth transposition over the 5 hour measurements of 23.5 ± 4.4 µ m and the other a transposition of 66.7 ± 13.7 µ m.The  rst category represented the sides with hyalinized PDL ( n = 7) and the latter the non-hyalinized sides ( n = 3).
For statistical analyses, the tooth transposition data for the hyalinized and the non-hyalinized group were split into  ve periods of 1 hour each.Separate analyses were performed, including or excluding the measurements of the  rst minute.The data were normally distributed (Kolmogorov -Smirnov), and based on a previous study ( Jónsdóttir et al. , 2006 ), data from different sides were considered to be independent.Therefore, time dependency of the amount and rate of transposition were analysed by analysis of variance and Tukey ' s multiple comparisons post hoc tests.The differences between the hyalinized and nonhyalinized sides at different time points were analysed with Mann -Whitney tests as Levene ' s tests showed unequal variances between the hyalinized and non-hyalinized group.

Results
Not all experimental sides had reached the hyalinization phase as shown by long-term time-displacement curves or post hoc histological evaluation ( Figure 2 ).The 5 hour measurement data of the hyalinized sides ( n = 7) and the non-hyalinized sides ( n = 3) are summarized in Figure 3A .

of 5
In both the hyalinized and the non-hyalinized groups, an initial rapid transposition was found, followed by a decreasing creep response during the remainder of the experimental period.The mean tooth transposition in the  rst 5 seconds for the hyalinized group was 9.9 ± 2.3 µ m and for the non-hyalinized group 31.3 ± 20.1 µ m ( P < 0.02 ; Figure 3B ).
The total transposition in the  rst hour was 20.5 ± 4.0 µ m in the hyalinized group and 55.5 ± 14.9 µ m in the nonhyalinized group ( P < 0.02 ; Figure 3A ).If the  rst minute was excluded, the mean remaining transposition in the  rst hour for the hyalinized group was 6.1 ± 4.0 µ m and for the non-hyalinized group 16.0 ± 5.7 µ m ( P = 0.03).The transposition rate for both groups decreased signi cantly ( P ≤ 0.002) during the subsequent 1 hour periods ( Figure 4 ).However, the transposition rate for the four subsequent 1 hour periods after the  rst hour showed no signi cant mutual differences within each of the groups.Comparison of the hyalinized and non-hyalinized groups showed signi cantly more creep in the non-hyalinized group for all 1 hour periods, except for the last hour ( Figure 4 ) .

Discussion
In the present study , a standardized orthodontic force of 100 cN was applied to the second premolars of dogs for 6 days.Thereafter, the PDL was allowed to relax for 24 hours.Then, a standardized force of 100 cN was again applied, and tooth transposition was measured over a 5 hour period to study the biomechanical behaviour of the PDL.The 24 hour relaxation period allowed for tooth transposition measurements under unstressed conditions, while the PDL morphology differs from normal.For the compression areas , this meant that in most cases , hyalinized tissue was present while the other cases showed non-hyalinized PDL.In the latter group, the PDL was devoid of principal  bres; it contained mainly loose connective tissue with randomly orientated thin collagen  bres as described in other studies ( Ren et al. , 2008 ;Lv et al. , 2009 ).The tension areas in the hyalinized as well as in the non-hyalinized group still contained principal  bres and blood vessels (Maltha, Jónsdóttir, unpublished observations ).This is in agreement with other studies ( Pilon et al. , 1996 ;Van Leeuwen et al. , 1999 ).
Under physiological conditions , teeth show a very rapid initial transposition upon force application that lasts for a few seconds to minutes ( Jónsdóttir et al. , 2006 ;Slomka et al. , 2008 ).This can be explained by a rapid  uid reallocation within the PDL ( Van Driel et al. , 2000 ).The initial phase is followed by a gradually decreasing creep movement that lasts for approximately 5 hours.This is probably related to the viscoelastic material properties of the PDL at the tension areas ( Van Driel et al. , 2000 ;Jónsdóttir et al. , 2006 ).In a previous investigation , it was shown that normal PDL can be considered as a poroviscoelastic  bre-reinforced material ( Jónsdóttir et al. , 2006 ).In the present study, the initial transposition in the presence of non-hyalinized PDL was approximately 39.5 ± 20 µ m in the  rst minute, which is higher than in the normal PDL (23 ± 22 µ m).Although this difference was not statistically signi cant, and the number of experimental sides was low, it suggests that the structural changes in the PDL at the compression areas, where the normal PDL is replaced by a loose connective tissue, allow for more rapid  uid reallocation.In the hyalinized group, however, the initial transposition in the  rst minute was signi cantly less than normal (14 ± 4 µ m), suggesting that  uid reallocation is inhibited by the hyalinized dense necrotic material ( Kurol and Owman-Moll, 1998 ;Melsen, 2001 ;Tomizuka et al. , 2007 ).
The total creep after the  rst minute in the non-hyalinized group was approximately 27 ± 7 µ m, which is comparable with that found in a previous study with a normal PDL (33 ± 16 µ m ; Jónsdóttir et al. , 2006 ).This suggests that the mechanical characteristics of the PDL in the tension areas remain the same during the  rst 6 days of orthodontic tooth movement.
The creep after the  rst minute in the hyalinized group was completely different.Although the histology of the PDL in the tension areas in this group was comparable to that in the non-hyalinized group, the creep was signi cantly reduced.Apparently, the high density of the hyalinized PDL and the absence of blood vessels in the compression areas inhibit creep movement.The  nal result is that during the 5 hour period, the transposition is signi cantly less in the hyalinized than in the non-hyalinized group.Assuming the normal width of the PDL to be about 0.2 mm, the evoked strain can be roughly estimated as 25 -30 per cent for normal 5 of 5 MECHANICAL BEHAVIOUR OF HYALINIZED PDL and non-hyalinized PDL and as 7 per cent for the hyalinized PDL.Consequently, the strain in the PDL and the alveolar bone is probably less if hyalinization is present.This might affect the signalling pathways that are responsible for the differentiation and/or recruitment of osteoclasts, osteoblast progenitors, and in ammatory cells ( Krishnan and Davidovitch, 2006 ;Henneman et al. , 2008 ).

Conclusions
The present study reports for the  rst time on the biomechanical behaviour of the PDL during orthodontic tooth movement in the presence or absence of hyalinization.It can be concluded that the null hypothesis that mechanical behaviour of the hyalinized PDL and non-hyalinized PDL is equal, is not supported by the present  ndings.This should be considered in future studies on the effects of force-induced tissue strain on cell and tissue reactions to orthodontic forces.Furthermore, these data can be used in FE analyses as these are only valid if realistic mechanical properties of the PDL are accurately incorporated ( Limbert et al. , 2003 ;Cattaneo et al. , 2005 ;Bosshardt et al. , 2008 ;Wakabayashi et al. , 2008 ).

Figure 2
Figure 2 Histological sections showing the period on tal ligament (PDL) of (a) non-hyalinized PDL at the pressure side, (b) hyalinized PDL at the pressure side, and (c) PDL at the tension side.B , alveolar bone; T , tooth .

Figure 3
Figure 3 Tooth transposition curves (means ± standard error) of the hyalinized and the non-hyalinized groups for (A) the 5 hour and (B)  rst minute of the measurements.

Figure 4
Figure 4Transposition rates (means ± standard error) for the hyalinized and non-hyalinized groups (excluding the  rst minute) over  ve subsequent 1 hour periods.The decrease in transposition rate from the  rst to the second period was signi cant for both groups ( P < 0.01).Asterisk indicates signi cant difference between groups at the indicated time points ( P < 0.05) .