Correlation between Gamma Glutamyl Transferase (GGT) and bone quality

Correlation between Gamma Glutamyl Transferase (GGT)

and bone quality

Phd Student Piolanti Nicola

Supervisor: Prof. Michele Lisanti

Tutors:

Prof. Aldo Paolicchi, Prof. Stefano Marchetti; Prof Vanna

Fierabracci; Phd. Lorenzo Andreani; Prof. Paolo Parchi

INDEX

1. SUMMARY ... 5

2.RATIONALE FOR THE STUDY ... 06

3.MATERIALS AND METHODS ... 12

3.1 RECLUTATION PHASE ... 12 Inclusion Criteria ... 12 Exclusion Criteria ... 12 3.2 PREOPERATIVE PHASE ... 16 3.3 INTRAOPERATIVE PHASE ... 16 4. RESULTS ... 18

4.1 BLOOD SAMPLE RESULTS ... 18

4.2 RISULTS FEMORAL NEC MICRO-CT ... 21

4.3 RESULTS ABOUT FEMORAL HEAD MORPHOLOGY ... 27

5. DISCUSSION ... 28

5.1 LITERATURE ... 28

5.2 CLINICAL STUDY ... 32

5.2.1 Analysis of blood tests ... 32

5.2.2 MicroCT analysis of femoral neck ... 32 5.2.3 Morphologic Analysis of Femoral Head ... 34 6. CONCLUSIONS ... 35 7. REFERENCES……...………...……….37

1. SUMMARY

The following clinical study has the principal aim to verify the possible correlation between GGT fraction (gamma-glutamyltransferase) values and bone quality.

We examined 93 Caucasian patients of both genders and aged between 60 and 85 years who were admitted to our Orthopaedic Unit with a diagnosis of hip osteoarthritis or proximal femoral fracture. Patients were divided in two main groups: Group A including patients who underwent total hip replacement because of hip arthritis and Group B including patients who underwent proximal femoral fracture ostheosynthesis by means of endomedullary nail, plate and screws or hip prosthesis. In all patients a preoperative blood sample was collected to obtain total and GGT fraction values. Furthermore, in all cases of hip prosthesis implantation, femoral head and neck were sampled. Femoral head was submitted to morphological analysis, while femoral neck was submitted to microCT analysis.

The results were analysed to assess the possible correlation between abnormal GGT fraction values and bone quality. The final results were compared with data from literature.

2.RATIONALE FOR THE STUDY

Gamma-glutamyltransferase (GGT) is an enzyme expressed on the membrane of all human cells. It is present in larger amount in tissues with secretory or absorptive function like liver, biliary tract and kidney. [1] GGT is implied in amino acid metabolism. In particular, it catalyses the transfer of a γ-glutamyl functional group from an amine to an acceptor molecule (Picture 1)[2].

The main physiological substrates of GGT are glutathione or γ-glutamylcysteinylglycine (GSH), Leukotriene C4 or LTC4 (a GSH conjugate), GSH conjugates produced by GSH-transferases and nitrosoglutathione. As GSH is primarily an antioxidant, even GGT has been thought to have an antioxidant role. Actually, GGT has proved to have a pro-oxidant action in some cases: for example in presence of copper [Cu++] or iron [Fe+++], GGT catalyses the formation of Reactive Oxygen Species (ROS) in the extracellular space (Figure 2).

As concerns GGT relationship with physiopathology, GGT elevation has long been associated with a major risk of cardiovascular pathologies, diabetes and metabolic syndrome[1]; only recently, a new correlation between GGT and bone mineralization has emerged. Another recent discovery has been the distinction of four different fractions of this enzyme: b-GGT (“big”), m-GGT (“medium”), s-GGT (“small”) and f-GGT (“free”). Each one of these fractions would have a specific role in different pathologies (e.g. liver and cardiovascular pathologies)[2;3;4;5;6]. On the basis of these new concepts, our study was designed to investigate the possible relationship between GGT fractions and bone remodelling.

2.2 BONE REMODELLING

The system of RANK (Receptor Activator of Nuclear Factor κB) and RANKL (Receptor Activator of Nuclear Factor κ B Ligand) is the basis of the phenomenon of bone remodelling and GGT is able to interact with it. So it seems useful to briefly describe this system.

Osteoclasts differentiation starts from the osteoblast secretion (under the stimulation of PTH [Parathyroid hormone], IL-11 [Interleukin-11] and 1α,25-dihydroxyvitamin D3 [7;8]) of M-CSF (macrophage colony-stimulating factor) (Picture 3).

RANKL is a protein expressed by osteoblasts and activated by lymphocytes T. It can bind RANK (localized on osteoclasts and monocytes) promoting differentiation and activation of osteoclasts that fuse together and provoke bone resorption [9] (Pictures 8-9); on the other hand Osteoprotegerin (OPG) is a receptorial protein that binds RANKL and inhibits RANK activation and the subsequent bone resorption.

Picture 3: Hormonal control of bone remodelling "Osteoclast differentiation and activation. William J. Boyle*, W. Scott

PIcture 8: RANK-RANKL meccanism of action

Picture 8: link between RANL-RANKL. From "Osteoclast differentiation and activation. William J. Boyle*, W. Scott Simonet & David L. Lacey
 "

In vivo studies in mice have evidenced that RANKL gene knockout animals lack osteoclasts, while OPG gene knockout mice present a severe osteoporosis [9]. Furthermore, it is interesting how the OPG/RANK/RANKL complex is influenced by inflammatory and/or immunologic stimuli, which in this way are able to increase bone loss and consequently the risk of fracture.

3. MATERIALS AND METHODS

From February 2014 to August 2015 we enrolled two groups of patients who were admitted to our Orthopaedic Unit because of orthopaedic or traumatologic problems.

Inclusion Criteria

-Group A: patients aged between 60 and 85 years suffering from primary osteoarthritis of the hip who need an intervention of total hip replacement; -Group B: patients aged between 60 and 85 years who have experienced a proximal femoral fragility fracture and need an intervention of total or partial hip replacement or reduction and synthesis of the fracture.

Exclusion Criteria

-Group A: anamnesis positive for liver diseases, alcohol abuse, osteoporosis (treated or not) or therapy with drugs that may have modified the quality of osseous tissue (for example chronic corticosteroid therapy)

-Group B: anamnesis positive for liver diseases or alcohol abuse

Before enrolment in our study, each patient signed an informed consent form (Picture 9). In addition, a data collection form was filled in for each patient (Picture 10).

Picture 9: informed consent form

Picture 10: data collection form

3.2 PREOPERATIVE PHASE

During the preoperative phase, a blood sample of 3 ml was collected from each patient in a test tube containing EDTA (ethylenediamine tetraacetic acid). Then, the test tubes were sent to our laboratory (Laboratorio di patologia Clinica dell’AOU Pisana) in order to analyse the GGT fractions. In laboratory the specimen were injected size exclusion chromatography column (Superose 6 HR 10/300 GL, GE Healthcare) and eluted at a constant speed of 0,5 ml/min from a mobile phase consisting of a sodium phosphate buffer (0.1 mM and pH 7.4) containing NaCl 0.2 M, EDTA 0.1 mM, glycilglycine (Gly-gly) 4.5 mM. The chromatography column was calibrated with proteins with a known molecular weight in order to establish the relationship between molecular weight and elution volume. In order to determine continuously GGT activity, by means of a T shaped connector, a solution containing a fluorogenic substrate for GGT [γGluAMC or L-Glutamic acid γ-(7-amido-4-methylcoumarin)] was continuously added with a flow of 0.1 ml/min. The subsequent enzymatic reaction proceeded for 4.5 minutes at pH 8.2 in a reaction coil with a volume of 2.6 ml kept at a constant temperature of 37°C in a water bath (Picture 11). As the acceptor of the reaction catalysed by GGT was included in the mobile phase, the presence of the enzyme was evidenced by the liberation of the fluorogenic substance 7-amido-4-methylcoumarin (AMC), like in the following reaction: γ-Glu-AMC + GlyGly → γ-Glu-GlyGly + AMC

The signal produced by the reaction was detected by means of a fluorescence detector operating at excitation/emission wavelength of 380/440 nm. With this technique it was possible to identify 4 different GGT fractions which elute at the following volumes: b-GGT 11.4 ml; m-GGT 15 ml; s-GGT 18.5 ml; f-GGT 21.2 ml. Under this reaction conditions, area under curve is proportional to GGT activity. Total area (between 10 and 25 mL elution volume) and GGT fraction area was calculated by the laboratory of the Clinical Physiology Institute (CNR of Pisa) to resolve overlapping peaks. The γGluAMC solution was prepared with a concentration of 4.5 mmol/L in ethanol 30% w/w containing NaOH 0.005 N; this solution was kept at -20°C, avoiding light exposition. The γGluAMC solution 4.5 mM was diluted, just before its use, in a buffer TRIS 0.25 M at pH 8.5 to obtain a γGluAMC concentration of 180 µM. The intensity of the fluorescence signal was expressed in arbitrary fluorescence units (f.u.) [3,10].

Thereafter, collected data were subjected to statistical analysis in order to determine the relationship between GGT fractions activity and personal, anamnestic and hematochemical parameters. The main statistical methods used were univariate and multivariate linear regression analysis in addition to calculation of parameters to describe the distribution of GGT fractions values. 3.3 INTRAOPERATIVE PHASE

Patients suffering from primary localized arthritis of the hip (Group A) underwent a total hip replacement. During surgery femoral head and neck were removed and preserved. The femoral head, fixed in formalin and subsequently embedded in paraffin, was sent to our laboratory of anatomical pathology (Anatomia Patologica III) where it was subjected to morphological analysis (on sections stained with haematoxylin and eosin) (Picture 12).

The femoral neck specimen, fixed in formalin, was analysed with MicroCT in the laboratories of Clinical Physiology Institute of CNR (Consiglio Nazionale delle Ricerche) of Pisa (Figure 13).

As regards patients with a fragility femoral fracture (Group B), a hip replacement was performed in case of a medial fracture (Figure 14) and femoral head and neck were sent to laboratories. On the other hand lateral fractures of the femoral neck were reduced and fixed by means of endomedullary nail, without any collection of bone tissue.

MicroCT analysis of the femoral neck was based on two main bone parameters: the bone volume fraction (Bone Volume/Total Volume or BV/TV) and the structure model index (SMI), which describes the shape of bone trabeculae (rod-like or plate-like). The values obtained were employed to stratify the patients on the basis of the calculation of quartiles distribution of BV/TV and SMI.

4.RESULTS

Between February 2014 and August 2015 we enrolled 93 patients, 61 females (f) and 32 males (m). 46 Patients in Group A (26f, 20m) and 47 in Group B (35f, 12m). Group A mean age was 74 years old (63 minimum, 83 maximum) and 83 in Group B (61 minimum, 85 maximum).

4.1 BLOOD SAMPLE RESULTS

With regard to blood samples results, we discovered a different distribution of GGT fractions in the two Groups (in the following graphics with the term “frattura” we refer to group B while “artrosi” is for group A).

As concerns b-GGT (graphic 1), this isoenzyme is on average slightly higher in Group B patients: Group B average value was 1,72 U/L and Group A average value 1,65 U/L.

Frattura Artrosi 0 5 10 15 20 25 Attività (U/L) bGGT Grafico 1

If we evaluate the m-GGT fraction (graphic 2), the values of this isoenzyme are only little discrepant between the two groups with a minimum increment in group B (0,39 U/L average value) compared to group A (0,345 U/L average value).

The values of s-GGT (graphic 3) are higher in Group B (5,07 U/L average value) than in Group A ( 4,01 U/L average value).

Frattura Artrosi 0 1 2 3 4 Attività (U/L) mGGT Graphic 2 Frattura Artrosi 0 10 20 30 40 50 Attività (U/L) sGGT Graphic 3

Unlike other fractions, f-GGT (graphic 4) is higher in group A (9,73 U/L average value) compared with group B (7,8 U/L average value).

With regard to total GGT values, these are higher in Group A (16, 44 U/L average value) compared with Group B (15,56 U/L average value).

The only datum with statistically significant difference between the two groups is f-GGT fraction (Table 1).

Frattura Artrosi 0 5 10 15 20 25 Attività (U/L) fGGT Graphic 4

4.2 FEMORAL NECK MICRO CT

With micro ct we evaluated the fraction of bone volume (BV/TV) and the structure model index (SMI). We stratified the patients in function of these two values related with GGT fractions.

On the basis of BV/TV, patients under the first quartile are considered ostheoporotic, while patients under the forth quartile are considered “normal” (Group A); as for as patients under the second and third quartile, we are not able to classify them on the basis of the value of BV/TV (Graphic 5).

BV/TV˚artrosi BV/TV˚fratture 0.00 0.05 0.10 0.15 0.20 ˚BV/TV˚artrosi_{˚BV/TV˚fratture} p=0.005˚(*) B V /T V ˚( m m ) Graphic 5

With regard to b-GGT fraction, this is higher in those patients with lower BV/TV (graphic 6) with a statistical significant result p=0,0393.

m-GGT (graphic 7) and s-GGT (graphic 8) are higher in patients with low BV/TV. I IV 0 2 4 6 8 Quartili BV/TV Attività (U/L) bGGT Graphic 6 I IV 0.0 0.5 1.0 1.5 2.0 Quartili BV/TV Attività (U/L) mGGT Graphic 7 Graphic 8 I IV 0 10 20 30 Quartili BV/TV Attività (U/L) sGGT

If we analyse f-GGT (graphic 9), we observe that the result changes: f-GGT is lower in patients with low BV/TV.

Total GGT is higher in patients with low BV/TV.

When we analyse the structure model index (SMI), a criterion that describes the shape of trabeculae (rod-like or plate-like), we find that in the patients of the 1° quarter (group A) there is a majority of plate like trabeculae (higher resistance), while in patients of the 4° quarter (group B) there is a majority of road like trabeculae.

b-GGT (graphic 10) is clearly higher in patients with road like SMI with statistically significant values p=0,0117.

I IV 0 5 10 15 20 25 Quartili BV/TV Attività (U/L) fGGT Graphic 9

.

m-GGT (graphic 11) and s-GGT fractions (graphic 12) are higher in patients with road like SMI. For m-GGT we find a statistically significant value p=0,0357.

As for B

V/TV, also for SMI f-GGT fraction values are countertrend (graphic 12) in fact it is higher in patients with plate-like SMI.

Total GGT are higher in patients with road-like SMI.

I IV 0 2 4 6 8 Quartili SMI Attività (U/L) bGGT Graphic 10 I IV 0.0 0.5 1.0 1.5 2.0 Quartili SMI Attività (U/L) mGGT I IV 0 10 20 30 Quartili SMI Attività (U/L) sGGT Graphic 11 Graphic 12

At the end we found statistically significant values only for b-GGT if we consider the BV/TV parameter, and for b-GGT and m-GGT if we consider SMI (table 2, 3).

Table 2: results and statistically significant values (red) between GGT fractions and BV/TV b e t w e e n 1 °

(osteoporotic patients) and 4° quarter (osteoarthritis patients) Graphic 12 I IV 0 5 10 15 20 25 Quartili SMI Attività (U/L) fGGT

between 1° (osteoporotic patients) and 4° quarter (osteoarthritis patients)

In pictures 21 and 22 we can see the differences in bone trabeculae from a micro-structural point of view between osteoarthritis patients (group A) and osteoporotic patients (group B).

Picture 14: bone trabeculae in a patient affected by osteoarthritis.

4.3 RESULTS ABOUT FEMORAL HEAD MORPHOLOGY

Out of a total of 93 patients, 32 femoral head were collected and analysed. The morphologic analysis conducted in the laboratories of Anatomia Patologica III of our Hospital confirmed the clinical suspicion. In fact Group A patients presented a remodelled osseous tissue with an irregular cortical bone and an altered articular cartilage, while Group B patients had an evident osteoporotic bone tissue with haemorrhagic areas and associated bone remodelling.

Picture 15: bone trabeculae in a patient affected by osteoporosis.

5. DISCUSSION

First of all, it is important to underline that in literature there are only few studies about the correlation between GGT and bone remodelling and none about the correlation between GGT fractions and bone. This fact was a negative factor because we found only few literature references but also an incentive to discover new correlations not yet studied.

5.1 LITERATURE

One of the first studies about the correlation between GGT and bone quality was written by di Shumpei Niida et al. in 2003 [11]. These Authors evaluated GGT effects on cultures of mouse bone marrow stromal cells. The result was an osteoclasts increase with a consequent bone loosening, linked not yet with GGT enzymatic activity but with its interaction with RANKL (as it happens with IL1). These data were confirmed by the fact that enzymatically inactive GGT (after reaction with acividin) was able to activate osteoclasts. This report was the first demonstration of a novel biological activity of GGT protein in a manner independent from its enzymatic activity.

Yutaro Asaba et al. in 2006 [12], proved the correlation between urinary GGT increase and augmented osteoclasts activity. These Authors found an increase of urinary excretion of GGT in osteoprotegerin (OPG)-deficient osteoporotic mice as well as in patients with postmenopausal osteoporosis (67–83 years of age); in both cases the urinary level decreased after treatment of patients or mice with alendronate, a selective inhibitor of bone resorption, concomitantly with a reduction in N-telopeptide (NTX) and deoxypyridinoline (DPD) (Picture 16)

In 2007 Kyoshi Hiramatsu et al.[13], analysing the effects of purified GGT on mice, studied even the morphology of osseous tissue by means of micro CT. They analysed tibial metaphysis and observed that GGT overexpression was associated with severe osteoporosis, reduced BV/TV, increased trabecukar space and altered SMI (Pucture 17-18).

Picture 16: tibiae microCT and tartrate-resistant acid phosphatase (TRAP) staining between tosteopetrotic op/op mice, op/op mice plus M-CSF, wild type mice “Urinary gamma-glutamyltransferase (GGT) as a potential marker of bone resorption.” Di Asaba Y et al.

Picture 17: three-dimensional images obtained by micro-CT of trabecular bone at the tibiae of wild-type (WT) and of GGT transgenic (Tg) mice from " Hiramatsu K, Asaba Y, Takeshita S, Nimura Y, Tatsumi S, Katagiri N, et al. Overexpression of gamma-glutamyltransferase in transgenic mice accelerates bone resorption and causes osteoporosis. Endocrinology. 2007; 148(6): 2708-15.

Picture 18: BV/TV, SMI, Tb.N (trabecular number) and Tb.Th (trabeculat thickness) in wild type (WT) mice and transgenic (Tg) mice from " Hiramatsu K, Asaba Y, Takeshita S, Nimura Y, Tatsumi S, Katagiri N, et al. Overexpression of gamma-glutamyltransferase in transgenic mice accelerates bone resorption and causes osteoporosis. Endocrinology. 2007; 148(6): 2708-15."

Unlike previous studies, in 2011 Canan tikiz et al. [14] denied a true correlation between urinary GGT and osteoporosis. They examined a group of menopausal women and divided them in 3 groups on the basis of their lumbar column T-score: osteoporotic, osteopenic and normal. Then, these patients underwent blood test for measuring the level of deoxypyridinoline, GGT, osteocalcin and osseous alkaline phosphatase. While osteocalcin, deoxypyridinoline and alkaline phosphatase increased in osteoporotic patients, GGT did not demonstrate a relevant increase in osteoporotic patients; in this way, they excluded the hypothesis that GGT could be a new marker of bone resorption.

A recent study of Beom-Jun Kim et al. [15] conducted on 16,036 men with age over 50 years old and with elevated GGT levels evidenced the possible correlation between high GGT levels and bone tissue quality. They dosed serum GGT levels in these patients and they calculated the percentage of fragility fractures in the following years. They found that patients with higher GGT levels presented a major risk of fracture. In contrast to the previous study, this result would reaffirm the link between GGT and osseous quality and its possible use as a marker of bone resorption.

According to some studies [2,16], the reason why GGT stimulates bone remodelling would be the presence of specific receptors for GGT on the surface of osteoblasts. The confirm of a cytokine-like action of GGT would lay in the structure of this protein: GGT molecule presents two disulphide bridges, one between Cys 49 and 73 and one between Cys 191 and 195 (placed following the CX3C scheme); these are structural elements that characterize a group of chemokines like neuroactin. Furthermore, the CXC3

5.2 CLINICAL STUDY

As concerns our study, it would have partially confirmed a possible correlation between GGT and bone quality.

5.2.1 Analysis of blood tests

The results of blood tests showed that the values of the different GGT fractions change in the two groups. Only f-GGT results are statistically significant (p=0,0005): Group B f-GGT are lower than Group A. Total GGT values are lower in Group B than in Group A, in contrast with what reported in literature. As regards b-GGT, m-GGT e s-GGT fractions, they do not vary between the two groups.

5.2.2 MicroCT analysis of femoral neck

Analysing microCT results, both BV/TV, which was lower in Group B, and SMI demonstrate an association between high b-GGT, m-GGT e s-GGT and total GGT values and osteoporosis (picture 19,20). It is important to underline that a statistical significance was found only BV/TV of b-GGT (p=0,0393) and for SMI of b-GGT (p=0,0117) and m-GGT (p=0,0357); in this way it would be possible to confirm a relationship between GGT and bone quality. Only f-GGT fraction has discordant trend, as it is higher in non-osteoporotic patients.

Piture 19: example of microTC group A

The results do not definitely confirm a clear correlation between GGT fractions and bone resorption when we consider blood test analysis, while results are more statistically significant for b-GGT fraction and partially for m-GGT fraction when we consider micro CT analysis.

5.2.3 Morphologic Analysis of Femoral Head

Finally, morphologic analysis evidenced how patients belonging to Group A (arthritis) present a particular bone structure with altered articular cartilage, cortical irregularities and bone remodelling. As concerns Group B (osteoporosis), patients present an osteoporotic osseous tissue with haemorrhagic areas and associated bone remodelling. This analysis confirms that patients with altered plasmatic GGT fractions values show an altered bone remodelling. Comparing our results with what reported in literature, we can affirm that our study, like most of the previous researches, evidences a correlation between altered GGT levels an bone quality, even if in our case f-GGT were decreased in osteoporotic patients, in contrast of what reported in literature. Then, another interesting aspect is the possible correlation between augmented f-GGT and arthritis. This could be the basis for a new study to verify a possible function of f-GGT as marker of this pathology.

6. CONCLUSIONS

The absence of literature regarding the relationship between GGT fractions and osteoporosis represented a limit of this study, but also a strong point and a stimulus. We would like to underline that our clinical study has been the first to examine the correlation between GGT fractions and bone quality, as other previously considered only total GGT values. Moreover, this prospective clinical study is the first to be conducted in vivo on human beings, while all the previous works were conducted in vitro or on in animal models.

After a careful analysis of our results and after a compare with data available in literature, we can draw the following conclusions.

Firstly, we can observe a relationship between GGT fractions levels (in particular b-GGT and m-GGT fractions) and bone quality. In fact, the levels of these fractions increase in osteoporotic fractions (in particular in patients with low BV/TV and rod-like SMI), confirming our assumption.

So, the results of this study would suggest that GGT could be used as a new serum marker of bone remodelling. If these results were confirmed by further studies with an enlargement of the number of patients examined, serum GGT level could become a new diagnostic instrument in order to make faster the diagnostic process for osteoporosis and to decrease costs for the patient and the society (for example reducing the number of radiographic exams).

Furthermore, high GGT levels could be considered as a risk factor for osteoporosis, so dosage of GGT values could become a valid instrument for screening, early diagnosis and follow up of this pathology.

our initial hypothesis and suggests a possible relationship between f-GGT and arthritis.

In conclusion, with this study we can confirm the initial hypothesis of correlation between GGT fractions and bone remodelling and we lay the foundations for new studies about the relationship between GGT fractions and arthritis.

7.

REFERENCES

1. Hanigan MH, Frierson HF. Immunohistochemical detection of gamma-glutamyl transpeptidase in normal human tissue. J Histochem Cytochem. 1996; 44(10): 1101-8.

2. Franzini M CA, Mammini C. γ-Glutammiltransferasi: biochimica clinica e fisiopatologia umana. Biochimica clinica. 33(1): 9-38.

3. Franzini M, Ottaviano V, Fierabracci V, Bramanti E, Zyw L, Barsacchi R, et al. Fractions of plasma gamma-glutamyltransferase in healthy individuals: reference values. Clinica chimica acta; international journal of clinical chemistry. 2008; 395(1-2): 188-9.

4. Belcastro E FM, Cianchetti S, Lorenzini E, Masotti S, Fierabacci V, Pucci A, Pompella A, Corti A. Monocytes/macrophages activation contributes to b-gamma-glutamyltransferase accumulation inside atherosclerotic plaques. Journal of Translational Medicine. 2015.

5. Franzini M CA, Fornaciari I, Balderi M, Torracca F, Lorenzini E, Baggiani A, Pompella A, Emdin M, Paolicchi A. Cultured human cells release soluble gamma-glutamyltransferase complexes corresponding to the plasma b-GGT. Biomarkers 2009.

6. Franzini M FI, Siciliano G, Voldi L, Ricci G, Marchi S, Gagliardi G, Baggiani A, Torracca F, Fierabacci V, Miccoli M, Pompella A, Emdin M, Paolicchi A. Serum gamma-glutamyltransferase fractions in myotonic dystrophy type 1: differences with healthy subjects and patients with liver disease. Clinical biochemistry. 2010.

7. Takahashi N, Akatsu T, Sasaki T, Nicholson GC, Moseley JM, Martin TJ, et al. Induction of calcitonin receptors by 1 alpha, 25-dihydroxyvitamin

8. Akatsu T, Takahashi N, Udagawa N, Sato K, Nagata N, Moseley JM, et al. Parathyroid hormone (PTH)-related protein is a potent stimulator of osteoclast-like multinucleated cell formation to the same extent as PTH in mouse marrow cultures. Endocrinology. 1989; 125(1): 20-7.

9. Boyle WJ, Simonet WS, Lacey DL. Osteoclast differentiation and activation. Nature. 2003; 423(6937): 337-42.

10. Franzini M, Bramanti E, Ottaviano V, Ghiri E, Scatena F, Barsacchi R, et al. A high performance gel filtration chromatography method for gamma-glutamyltransferase fraction analysis. Anal Biochem. 2008; 374(1): 1-6.

11. Niida S, Kawahara M, Ishizuka Y, Ikeda Y, Kondo T, Hibi T, et al. Gamma-glutamyltranspeptidase stimulates receptor activator of nuclear factor-kappaB ligand expression independent of its enzymatic activity and serves as a pathological bone-resorbing factor. The Journal of biological chemistry. 2004; 279(7): 5752-6.

12. Asaba Y, Hiramatsu K, Matsui Y, Harada A, Nimura Y, Katagiri N, et al. Urinary gamma-glutamyltransferase (GGT) as a potential marker of bone resorption. Bone. 2006; 39(6): 1276-82.

13. Hiramatsu K, Asaba Y, Takeshita S, Nimura Y, Tatsumi S, Katagiri N, et al. Overexpression of gamma-glutamyltransferase in transgenic mice accelerates bone resorption and causes osteoporosis. Endocrinology. 2007;

148(6): 2708-15.

14. Canan Tıkız CU, Fatma Taneli, Ebru Yiğit Acar, Gül Gümüşer, Gönül Dinç Horasan. Can Urinary Gamma Glutamyl Transferase be Used as a Bone Resorption Marker in Postmenopausal Osteoporosis? Turkish Journal of Biochemistry. 2011; 36(2): 154-9.

15. Kim BJ, Baek S, Ahn SH, Kim SH, Jo MW, Bae SJ, et al. A higher serum gamma-glutamyl transferase level could be associated with an

16. Kinlough CL PP, Bruns JB, et al. . Gamma- glutamyltranspeptidase: disulfide bridges, propeptide cleavage, and activation in the endoplasmic reticulum. Methods in enzymology. 2005.