A preview of this full-text is provided by SAGE Publications Inc.
Content available from Journal of Applied Biomaterials and Biomechanics
This content is subject to copyright.
INTRODUCTION
The improved understanding in material-cell inter-
actions results in a high degree of predictability in
the clinical success of biomaterials used for dental
implants (1, 2). Titanium has been established as
the first choice material for endosseous implants
because of its chemical-physical properties and its
biocompatibility (3, 4).
Restorations in the anterior esthetic zone present
significant challenges in both the surgical and pros-
thetic phases of implant dentistry and many types
of implants require transmucosal abutments to re-
tain implant restorations. In these cases, the gray
color of the titanium is transmitted through the
peri-implant tissues causing patient discomfort.
The use of ceramic abutments in methods with sub-
merged implants allows the minimization of the
gray color associated with metal components. Sub-
merged implants, on the other hand, evidenced
limits connected to the operative phases: a second
surgical step, longer clinical times, and the irre-
versibility of the prosthetic structure. Therefore,
some previous studies compared non-submerged
healing implants, also defined as one-stage im-
plants or transmucosal healing implants, with sub-
merged healing implants, also defined as two-stage
implants. These studies, found no differences in
Journal of Applied Biomaterials &Biomechanics 2004; 2: 143-150
In vitro and in vivo follow-up of titanium
transmucosal implants with a zirconia collar
A.E. BIANCHI1, M. BOSETTI2, G. DOLCI Jr1, M.T. SBERNA1, F. SANFILIPPO1, M. CANNAS2
1 Cattedra di Clinica Odontoiatrica, Ateneo Vita e Salute, San Raffaele Hospital, Milano - Italy
2 Department of Medical Sciences, University of Eastern Piedmont, Novara - Italy
ABSTRACT:The advantages of transmucosal healing implants with a bioactive zirconia collar as a support for partially fixed
prosthodontic restorations are optimal peri-implant marginal tissue sealing, reduction in plaque accumulation and satisfacto-
ry esthetic results. The zirconia used in this study evidenced not only optimal clinical performances, but also good biocompati-
bility. The results from this study demonstrated that zirconia coating enhances fibroblasts and osteoblast-like cell adhesion,
spreading and proliferation, favoring microscopic tissue/cell in-growth and clinical implant fixation improvement. From clin-
ical analysis, it emerged that the treatment group obtained better scores in every peri-implant parameter. This evidence attests
faster stabilization of soft and hard tissues around both the transmucosal zirconia collar and at the crestal level of the implant.
A reduced plaque accumulation around the implant with zirconia collar could provide a better peri-implant microbiological en-
vironment by allowing the soft tissues expression of optimal sealing and good bone adaptation to loading. From these clinical
and radiographic comparative analyzes, it emerged that in the treatment group the mean values were always similarly low. A
rapid stabilization of both hard and soft peri-implant tissues was documented in the 1st yr. In the treatment group, there was
the formation of stable tissue sealing the zirconia collar, which could preserve mucosal and bone levels. In conclusion, 2-yr clin-
ical results demonstrated that implants supporting fixed restorations using transmucosal healing implants with a zirconia col-
lar appeared a valid method, reporting 100% implant survival rates. Moreover, in vivo results obtained using strict parame-
ters to assess the peri-implant status affirmed that a zirconia collar offers excellent biological acceptance. Our preliminary in
vitro results statistically evidenced increased fibroblast and osteoblast adhesion and proliferation to zirconia compared to tita-
nium, and an index of enhanced material integration with bone and soft tissue cells. (Journal of Applied Biomaterials &Bio-
mechanics 2004; 2: 143-50)
KEY WORDS: Osteoblast like cells, Osteointegration, Zirconia
Received 04/01/04; Revised 25/01/04; Accepted 04/11/04
1722-6899/143-08$15.00/0
© Società Italiana Biomateriali
the long-term prognosis of endosseous dental im-
plants (5, 6), promising important clinical advan-
tages for the one-stage surgery technique.
The use of a new implant with a white transmucos-
al ceramic collar allowed the combination of the
advantages offered by transmucosal implant meth-
ods with a one-stage technique with natural trans-
parency and increased biocompatibility at the trans-
mucosal portion.
Zirconia is an advanced ceramic used in the bone
medical field (7) due to its important mechanical
properties and tissue biocompatibility. In addition,
zirconia has been shown to reduce bacterial adhe-
sion (8) and plaque accumulation (9).
In this study, we evaluated in vitro the biocompati-
bility of titanium compared to zirconia coated tita-
nium by studying fibroblasts and human osteoblast-
like cell adhesion and proliferation. In addition, we
performed preliminary clinical evaluations of trans-
mucosal healing implants with and without a zirco-
nia collar.
MATERIALS AND METHODS
Materials
Two types of solid screw transmucosal healing im-
plants with rough endosseous surfaces (Z1 im-
plants, T.B.R.®ide@, Sudimplant, Toulouse,
France) were used. For both types, the endosseous
surface was titanium, while the transmucosal collar
was titanium or titanium covered by a zirconia ring.
The ceramic ring was a yttria stabilized medical
grade zirconium dioxide.
The implants used for the clinical studies, in rela-
tion to bone height disposal, had an intraosseous
length ranging from 10.5 to 15.5 mm, and concern-
ing the bone width disposal, had an intraosseous di-
ameter from 3.5 to 5 mm. For in vitro tests, titanium
and zirconia disks of 0.4 cm diameter were used. Pri-
or to use, the coated and uncoated materials were
sterilized for 2 hr in dry heat at 160 °C.
In vitro evaluations
Cell behavior on the materials tested was studied us-
ing human fibroblasts (MRC5, ATCC cell-line) and
human primary osteoblast-like cells obtained by en-
zymatic isolation from trabecular fragments of
adult human bone removed during surgery, and
treated as described previously (10). Briefly, parti-
cles of 3-5 mm, after treatment with bacterial colla-
genase, were plated in 90 mm tissue culture dishes
and cultured at 37 °C in 95% air/5%CO2in 10 ml
ISCOVE’s supplemented with 20% FBS, 50 U/ml
penicillin, 15 µg/ml streptomycin and 2 mM gluta-
mine. Cell outgrowths from the bone fragments ap-
peared within 1 week and formed a confluent
monolayer at 3-4 weeks. The isolated bone cells
were characterized including osteoblastic morphol-
ogy, alkaline phosphatase expression and hormone
responsiveness (parathyroid hormone, 1,25(OH)2D3).
Fibroblasts and osteoblasts were used at a cell den-
sity of 1 x 104cells/cm2.
Cell adhesion and proliferation results were com-
pared to that obtained on a polymeric substrate
used as a control (Thermanox®slides, Nunc, Mi-
lano, Italy) known to induce cell adhesion (11).
As the materials tested were unsuitable for normal
transmitted light microscopy, we observed cell mor-
phology and distribution with fluorescence mi-
croscopy. After 6 hr (cell adhesion study) and 4
days (cell proliferation study), fibroblasts and os-
teoblast-like cells on the materials tested were
rinsed in phosphate buffered saline, fixed for 20
min at 60 °C and stained for 5 min in a 0.025% acri-
dine orange solution, a nucleic acid staining (12).
Cell morphology and the number on each material
were evaluated using a fluorescent microscope
Aristoplan (Leitz Leica, Milano, Italy). Cell number
was evaluated on a surface of 0.1715 mm2using 25x
microscope magnification and results are reported
as mean ± standard deviation of 30 fields obtained
from three different experiments.
Clinical evaluation
A 2-yr randomized study was performed from 2000-
2002. Twenty patients were admitted to the trial,
and 44 implants were placed; 29 implants had zir-
conia collars (Z1 implants, T.B.R. ide@, Sudim-
plant, Toulouse, France) and 15 implants had stan-
dard titanium collars. Thirteen patients received
only the implants with a zirconia collar, while five
patients received implants with the same shape but
with a titanium collar. The remaining two patients
were implanted with both implants. Data from pa-
tients who received implants with a zirconia collar
were processed as the treatment group, while data
collected from patients who received implants with
titanium collar were analyzed as the control group.
The number of implants placed in the maxilla was
24, and 20 implants were placed in the mandible.
The implants were used to support partially fixed
prosthesis and underwent loading at 3-4 months af-
ter their placement.
The cumulative survival rate was calculated accord-
ing to the method described by Cutler and Ederer
(13). The establishment of successful outcomes re-
144
In vitro and in vivo follow-up of titanium transmucosal implants with a zirconia collar
quired the periodical evaluation of specific para-
meters, such as the stability of peri-implant crestal
bone levels and the health and stability of peri-im-
plant soft tissue. According to Brägger et al (14),
crestal bone levels were evaluated calculating the
linear distance implant-shoulder to the bone
(DIB), measured with periapical radiographs on
mesial and distal sites (Fig. 1). When the DIB value
is <3.5 mm, peri-implant bone is considered stable;
>3.5 mm bone crest resorption has occurred.
The soft tissue status was evaluated clinically by the
method proposed by Mombelli et al (15) for trans-
mucosal healing implants and the following mea-
sures were recorded (Fig. 1).
Plaque index (PLI): This index is used to measure
the level of patient oral hygiene performance; it
confirms the quantity of bacterial deposits around
the implant emerging from soft tissues. Value 0 =
no plaque deposit, value 1 = small plaque deposit,
value 2 = large plaque deposits.
Bleeding on probing (BOP): This index is used to
measure the inflammation level of the mucosal tis-
sues in response to peri-implant sulcus probing.
Value 0 = no bleeding occurred during the prob-
ing, value 1 = small mucosal bleeding occurred, val-
ue 2 = significant bleeding occurred and value 3 =
spontaneous bleeding, even if the mucosal sulcus
has not been probed.
Probing depth (PD): This index measures (in mm)
the mucosal sulcus depth around the implant. A
small thin calibrated linear probe is inserted in the
peri-implant sulcus with a pressure of only 25 g. As
a starting point for the measurement, the free mar-
gin of the peri-implant mucosa is used, and an end-
point is reached when the resistance of the soft tis-
sue does not allow further probing. When the PD
value is <3 mm, peri-implant tissues are considered
healthy; >3 mm mucosal pathology has occurred.
Probing attachment level (PAL) is related to the im-
plant shoulder. An analogous procedure of PD
measurement, but as a starting point the platform
of the implants is used. When the PAL value is <2.5
mm, peri-implant tissues are considered healthy;
>2.5 mm mucosal pathology and bone resorption
have occurred.
Regular follow-ups from 4-6 months were established
based on individual oral hygiene performance, but
clinical data considered in this study were only those
recorded annually.
Statistical analysis
Statistical analysis of the data was carried out using
SPSS for Windows software. The Student’s t-test for
independent data was performed to compare ad-
hesion and proliferation results. A p value was ob-
tained from the ANOVA table; the conventional
0.05 level was considered statistically significant.
RESULTS
In vitro results
As shown in Figure 2, fibroblasts and osteoblast-like
cells evidenced statistically higher cell adhesion
when cultured on zirconia compared to the con-
trols (Thermanox®slides) and compared to the un-
coated titanium material.
145
Bianchi et al
Fig. 1 - Illustration depicting reference point and distances of clin-
ical and radiographic measurements: PD (probing depth), PAL
(probing attachment level, related to implant shoulder), DIB
(distance implant shoulder/first implant bone contact).
Fig. 2 - Fibroblasts and osteoblast-like cells quantification at 6 hr
cell incubation (adhesion test) on titanium (Ti), zirconia coated
titanium (Ti-Z1) and on control Thermanox®slide (CTR). Cell
number was the mean of 20 measurements in three experiments
(n=60) and was referred to a surface area of 0.1715 mm2.
* p<0.05 with respect to controls; ° p<0.05 with respect to Ti.
*
*
*°
*°
Figure 3 shows the results of fibroblasts and os-
teoblast-like cells at 4 days proliferation on the ma-
terials tested. Results evidenced increased cell num-
bers on the titanium and zirconia coated titanium
with respect to the adhesion experiments, compa-
rable to that obtained at 6 hr.
Human fibroblasts (Fig. 4a) and osteoblast-like cells
(Fig. 4b) cultured on the surfaces of the two mate-
rials and observed at 6 hr incubation showed good
fibroblast and osteoblast-like cell spreading on zir-
conia coated titanium, comparable to the control
Thermanox®wells.
Clinical results
During the 2-yr observation, no implant was lost ei-
ther in the treatment group or in the control
group. In the follow-up period, all implants showed
no mobility denoting clinical signs of stable os-
seointegration; therefore, an implant survival rate
146
In vitro and in vivo follow-up of titanium transmucosal implants with a zirconia collar
Fig. 3 - Fibroblasts and osteoblast-like cells quantification at 4 days
cell incubation (adhesion test) on titanium (Ti), zirconia coated ti-
tanium (Ti-Z1) and on control Thermanox®slide (CTR). Cell num-
ber was the mean of 20 measurements in three experiments (n=60)
and was referred to a surface area of 0.1715 mm2. * p<0.05 with
respect to controls; ° p<0.05 with respect to Ti.
Fig. 4 - Morphology of fibroblasts (a) and osteoblast-like cells (b) adhesion on cell culture dish, titanium and zirconia coated titanium.
Materials were incubated for 6 hr with fibroblasts and human osteoblast-like cells. Samples were examined by fluorescence microscopy
at 250
magnification.
Fibroblast adhesion
Osteoblast adhesion
a
b
CTR
CTR
*
*
*°
*°
of 100% was confirmed. Table I depicts the longevi-
ty of the implants, showing interval and cumulative
survival rates both for the treatment and control
groups.
Many patients exhibited good oral hygiene perfor-
mance, and in most patients, crowns or partially
fixed dentures were free from plaque or calculus
deposits. Recorded peri-implant parameters
demonstrated a PLI of 0 at 64.5% of all sites, and 1
at 29.5% and 2 at 6% of the other sites, respective-
ly. During the observation period, analogous mean
results were recorded in both groups, with a small
percentage increase in the scores for the treatment
group (Tab. II). Table II reports the differences in
the mean values of BOP at the final evaluation. The
BOP scores in the treatment group appeared lower
as compared to the control group.
In both procedures, a slight increase in PD mean
values, from 2.5 mm to approximately 3.5 mm, was
encountered during the 1st yr of load application
(Tab. III). Taking into consideration that the
healthy conditions of the soft tissue does not admit
a PD up to 3 mm, an analysis of PD for mean scores
>3 mm was required. The data revealed better
mean values of PD scores in the treatment group
during the 2-yr interval suggesting a faster stabiliza-
147
Bianchi et al
TABLE I - LIFE TABLE ANALYSIS FOR 44 IMPLANTS SURVIVAL
Interval N° of implants Drop-outs Implants Failure Survival Cumulative
months at start during under risk during rate within survival
of interval interval interval period (%) rate (%)
Total Treat* Cont** Total Treat Cont Total Treat Cont Total Treat Cont Total Treat Cont Total Treat Cont
12 44 29 15 2 2 0 42 27 15 0 0 0 100 100 100 100 100 100
24 9 5 4 0 0 0 9 5 4 0 0 0 100 100 100 100 100 100
* Treatment group ** Control group
TABLE II - PLI AND BOP INDEXES, PERCENTAGE OF MEAN VALUES AT LAST CONTROL
Collar Value 0 Value 1 Value 2
PLI Zirconia 72 24 4
Titanium 50 40 10
BOP Zirconia 88.9 8.3 2.8
Titanium 53.3 36.7 10
Mean values of PLI and BOP were significantly lower in treatment group
TABLE III - PERI-IMPLANT MEASURES RELATED TO OBSERVATION PERIOD, MEAN VALUES IN MILLIMETERS
Groups Time period Time period Time period Time period
0 months 6 months 12 months 24 months
PD Treatment 2.3 2.8 3 2.5
Control 2.8 3.2 3.4 3.3
PAL Treatment 1.8 2 2.1 0.5
Control 2.2 2.2 2.3 2.6
DIB Treatment 3 3 3.2 3
Control 3 3.2 3.6 3.4
Mean peri-implant values were lower in treatment group during interval
tion of peri-implant tissues (Tab. IV).
The trend in PAL scores was similar to PD scores
(Tab. III). For PAL scores it was also necessary to
confirm the percentage of healthy sites. Therefore,
the percentage of sites probed was calculated for
each single interval group with values up to 2.5 mm
(Tab. IV). The trend in this parameter appeared
analogous to PD for scores up to 3 mm.
Biologically, the distance between DIB crest should
be approximately 3.5 mm for the implants used in
this study. A linear measure of DIB ≤3.5 mm was as-
sumed as optimal. DIB scores >3.5 mm were then
collected and processed for comparative analysis.
The data collected in Table IV demonstrated the
different trend in the values among the treatment
group and the control group during the 1st yr. The
percentages of DIB scores >3.5 mm were integrated
with the percentage of PAL mean scores >2.5 mm
and with PD scores >3 mm.
DISCUSSION AND CONCLUSIONS
In the past 25 yrs, numerous in vivo studies have
demonstrated that non-submerged titanium im-
plants achieve osseointegration as predictable as
that of submerged titanium implants (4). This ob-
servation was confirmed in prospective clinical
studies. The studies demonstrated success rates well
above 90%. In summary, the non-submerged ap-
proach is a true alternative to the original healing
modality with submerged titanium implants. The
non-submerged approach offers several clinical ad-
vantages: 1) the avoidance of a second surgical pro-
cedure and less chair time per patient, resulting in
overall reduced treatment cost; 2) the lack of mi-
crogap at the bone crest level, leading to less crestal
bone during healing and resulting in a more favor-
able crown-to-implant length ratio; and 3) a simpli-
fied prosthetic procedure, presenting an ideal basis
148
In vitro and in vivo follow-up of titanium transmucosal implants with a zirconia collar
TABLE IV - COMPARATIVE ANALYSIS OF PERCENTAGE OF PERI-IMPLANT MEAN SCORES RECORDED WITH PD
> 3 mm, PAL > 2.5 mm AND DIB > 3.5 mm
Groups Time period Time period Time period Time period
0 months 6 months 12 months 24 months
PD > 3 mm 0 5.2 10.3 0
Treatment PAL > 2.5 mm 0 5.2 11.1 0
DIB > 3.5 mm 0 1.7 5.6 0
PD > 3 mm 0 60 53.3 50
Control PAL > 2.5 mm 0 30 56.7 50
DIB > 3.5 mm 0 6.7 20 0
More stable peri-implant status was attested in treatment group
Fig. 5 - Comparative analysis depicting the trend of peri-implant tissues in the treatment group (a) and the control group (b). Mean
values were calculated with PAL >2.5 mm, PD >3 mm and DIB >3.5 mm. Mean scores attested more stable peri-implant status in treat-
ment group, during 2 yrs of analysis.
ab
for cemented implant restorations. Due to these
significant clinical advantages, the non-submerged
approach will become more important in implant
dentistry in the near future, particularly in implant
sites without esthetic priority. The creation of a new
implant with a white transmucosal ceramic collar al-
lowed the combination of the advantages offered
by a transmucosal implant method with natural
transparency at the transmucosal level widening ap-
plications in restorations in the anterior esthetic
zone.
The advantages of transmucosal healing implants
with a bioactive zirconia collar as a support of par-
tially fixed prosthodontic restorations are optimal
peri-implant marginal tissue sealing, reduction in
plaque accumulation and satisfactory esthetic re-
sults. Zirconia is a widely studied material that has
evidenced good mechanical performances (7) with
reduced bacterial colonization (8) and reduced
plaque accumulation (9). In addition, the zirconia
used in this study evidenced not only optimal clini-
cal performances but also good biocompatibility. In
vitro tests demonstrated that zirconia coated titani-
um, compared to titanium, has a better compatibil-
ity related to the aspect considered. Results from
this study show that zirconia coating enhances fi-
broblasts and, particularly, osteoblast-like cell adhe-
sion, spreading and proliferation, favoring micro-
scopic tissue/cell in growth and clinical implant fix-
ation improvement.
From clinical analysis, it emerged that the treat-
ment group obtained better scores in every peri-im-
plant parameter. This evidence attests faster stabi-
lization of soft and hard tissues around both the
transmucosal zirconia collar and at the crestal level
of the implant (Fig. 5a). The same parameters eval-
uated for implants with titanium collars demon-
strated, instead, a major sensitivity of the peri-im-
plant tissues to the modifications of the biological
and biomechanical environment once the implant
supporting rehabilitation started working (Fig. 5b).
The association of factors such as plaque accumula-
tion and load can perhaps play an important role in
determining this tissue response. Reduced plaque
accumulation around the implant with a zirconia
collar could provide a better peri-implant microbi-
ological environment by allowing the soft tissue ex-
pression of optimal sealing and good bone adapta-
tion to loading. From these clinical and radi-
ographic comparative analyzes, it emerged that in
the treatment group the mean values were always
similarly low. A rapid stabilization of both hard and
soft peri-implant tissues was documented in the 1st
yr. In the treatment group, there was the formation
of stable tissue sealing the zirconia collar, which
could preserve mucosal and bone levels.
In conclusion, 2-yr clinical results demonstrated
that implants supporting fixed restorations using
transmucosal healing implants with zirconia collars
appeared as a valid method, reporting 100% im-
plant survival rates. In addition, in vivo results ob-
tained using strict parameters to assess the peri-im-
plant status confirmed that a zirconia collar offers
excellent biological acceptance according to our
preliminary in vitro results, which evidenced statisti-
cally increased fibroblast and osteoblast adhesion
and proliferation on zirconia coated titanium com-
pared to the uncoated form, and an index of en-
hanced material integration with bone and soft tis-
sue cells.
Address for correspondence:
Dr. Andrea E. Bianchi
Servizio di Odontoiatria,
Ateneo Vita e Salute,
Ospedale San Raffaele
Via Olgettina, 60
20132 Milano - Italy
bifasa@fastwebnet.it
149
Bianchi et al
REFERENCES
1. Ratner B. Replacing and renewing: synthetic materi-
als, biomimetics and tissue engineering in implant
dentistry. J Dent Educ 2001; 65: 1340-7.
2. Keller J. Physical and biological characteristics of im-
plant materials. Adv Dent Res 1999; 13: 5-7.
3. Akagawa Y, Abe Y. Titanium: the ultimate solution or an
evolutionary step? Int J Prosthodont 2003; 16: 47-51.
4. Buser D, Mericske-Stern R, Dula K, Lang NP. Clinical
experience with one-stage, non-submerged dental
implants. Adv Dent Res 1999; 13: 153-61.
5. Fiorellini JP, Buser D, Paquette DW, Williams RC,
Haghighi D. A radiographic evaluation of bone heal-
ing around submerged and non submerged dental
implants in beagle dog. J Periodontol 1999; 70: 248-
54.
6. Zechner W, Kneissel M, Kim S, Ulm C, Watzek G,
Plenk H. Histomorphometrical and clinical compar-
ison of submerged and nonsubmerged implants sub-
jected to experimental peri-implantitis in dog. Clin
Oral Implants Res 2004; 15: 23-33.
7. Hulbert SF. The use of alumina and zirconia in sur-
gical implants. In: Hench L, Wilson J eds. An intro-
duction to bioceramics, advances series in ceramics.
New York: World Scientific, 1993; 25-9.
8. Rimondini L, Cerroni L, Carrassi A, Torricelli P. Bac-
terial colonization of zirconia ceramic surfaces: an in
vitro and in vivo study. Int J Oral Maxillofac Implants
2002; 6: 793-8.
9. Grossner-Schreiber B, Griepentrog M, Haustein I, et
al. Plaque formation on surface modified dental im-
plants. An in vitro study. Clin Oral Implants Res 2001;
6: 543-51.
10. Bosetti M, Vernè E, Ferraris M, Ravaglioli A, Cannas
M. In vitro characterisation of zirconia coated by
bioactive glass. Biomaterials 2001; 22: 987-94.
11. Ehlers E, Behrens P, Wunsch L, Kuhnel W, Russlies
M. Effects of hyaluronic acid on the morphology and
proliferation of human chondrocytes in primary cell
culture. Ann Anat 2001; 183: 13-7.
12. Kepner RL, Pratt JR. The use of fluorocromes for di-
rect enumeration of total bacteria in environmental
samples: past and present. Microbiol Rev 1994; 58:
603-15.
13. Cutler SJ, Ederer F. Maximum utilization of the life
table method in analyzing survival. J Chronic Dis
1958; 6: 699-712.
14. Brägger U, Häfeli U, Huber B, Hämmerle CHF,
Lang NP. Evaluation of postsurgical crestal bone lev-
els adjacent to non-submerged dental implants. Clin
Oral Implants Res 1998; 9: 218-24.
15. Mombelli A, Van Osten, MAC, Schürch Ejr, Lang NP.
The microbiota associated with successful or failing
osseointegrated titanium implants. Oral Microbiol
Immunol 1987; 2: 145-51.
150
In vitro and in vivo follow-up of titanium transmucosal implants with a zirconia collar
Content uploaded by Michela Bosetti
Author content
All content in this area was uploaded by Michela Bosetti
Content may be subject to copyright.