Correlations between different protein species of oral rinse MMP‐8 and patient‐related factors

Abstract Objectives The aim of this study is to examine correlations between different oral rinse matrix metalloproteinase (MMP)‐8 protein species in western blot (WB) analysis, quantitative MMP‐8 measurements, and patient‐related factors. Elevated activated MMP‐8 (aMMP‐8) associate with periodontitis and a diagnostic point‐of‐care technology has been developed based on aMMP‐8. In WB, different MMP‐8 protein species can be analyzed. Relative abundancy of fragmented 20–25 kDa forms in WB has been associated with and reflects MMP‐8 activation and related fragmentation and elevated quantitative aMMP‐8 measurements. Material and Methods A random sample of 192 participants from a periodontal disease screening study was used for this study. Oral rinse samples for biomarker analyses were collected before clinical periodontal examinations. aMMP‐8 immunofluorometric (IFMA) and WB analysis (utilizing the same monoclonal antibody, 8708), polymorphonuclear leukocyte (PMN) elastase activity test and tissue inhibitor of metalloproteinases (TIMP)‐1 ELISA levels were performed from the oral rinse samples. Distinct MMP‐8 protein species were differentiated in the WB analysis. Principal component (PC) analysis was conducted to explore correlation patterns between the different species. Adjusted correlation analysis between the extracted PCs of WB and aMMP‐8 IFMA levels and multilevel regression analysis were conducted to explore if the other periodontal disease‐related biomarkers and clinical surrogate measures and patient‐related factors are co‐variating with the extracted components. Results Distinct correlation patterns between the MMP‐8 protein species were observed. The first four PCs explained 89% of the whole variance in PC analysis. Statistically significant correlation (p < 0.05) were observed as follows: PC1 positively with 21 kDa (r = .69) and 25 kDa fragments (r = .55) and negatively with 150 kDa complexes (r = −.46). PC2 correlated with 45 (r = .70) and 55 kDa (r = .65) activated forms, PC3 with 70–80 kDa latent proforms (r = .63) and 90–100 kDa complexes (r = .67), and PC4 with 35 kDa fragments (r = .81). There were significant correlations between quantitative (IFMA) aMMP‐8 measurements and PC1 (p < 0.001), PC2 (<0.05) and PC3 (<0.05) but not with PC4. In multilevel regression models age, PMN elastase activity, TIMP‐1 levels, and a number of 4–5 mm periodontal pockets were associated with PC1, nonsmoking with PC2, age and PMN elastase activity with PC3, and age and smoking with PC4. Conclusions Relative abundancy of fragmented 21–25 kDa protein species was correlated with the quantitative aMMP‐8 (IFMA) measurements, which is in line with previous results. Different patient‐related factors (smoking, age, proteolytic activity) may modify the formation of different MMP‐8 protein species in oral rinse samples and may cause variability in quantitative aMMP‐8 measurement.

MMP regulation is a complex process involving synthesis, secretion, activation, and inhibition (Romanelli et al., 1999).When neutrophils are recruited to a site of inflammation, they release latent MMP-8 stored in specific granules.Removal of the prodomain by enzymes of host or microbial origin results in a reduction of the molecular mass of the latent proMMPs (DeCarlo et al., 1997;Ding et al., 1995;Romanelli et al., 1999;Sorsa et al., 1992Sorsa et al., , 1995)).Different proteolytic enzymes cleave at different sites in the proenzyme domain, generating different sizes of active and fragmented enzymes with different levels of enzyme activity (Romanelli et al., 1999).There was some discrepancy in the western blot (WB) molecular weights of latent and active MMP-8 described in the early literature with values ranging from 85 kDa (Knauper et al., 1990) to 58 kDa (Uitto et al., 1990;van Wart, 1992) reported from polymorphonuclear leukocytes (PMN) preparations or from gingival crevicular fluid (GCF) samples.Different sample materials, preparations, and procedures may, at least partly, explain the differences in molecular sizes and forms, and the antibodies used in immunoblot may also give variant results.
Bacterial proteinases present in supra-and subgingival plaque can activate and process the PMN-type MMP-8 to less glycosylated 40-60 kDa forms.Species ≤30 kDa are regarded as fragments (Kiili et al., 2002).Romanelli et al. (1999) found these processed forms of MMP-8 from GCF, which was sampled by oral rinse method from periodontitis patients.They considered the smaller MMP-8 forms (around 60 kDa) as biologically important and hypothesized, that these species may be efficiently activated or even superactivated fragmented species.
The aim of this study is to further explore correlations between different MMP-8 protein species in WB analysis and quantitative IFMA measurements, patient-related factors, and periodontal disease surrogates.

| Patients
The clinical data of this study was originally designed and collected for a study testing the utility of neutrophil elastase activity test for screening periodontitis patients.Randomly selected 214 patients, who sought general dentists' treatment in public oral health services in the cities of Helsinki and Vantaa in Finland, were included in this cross-sectional study.Patients in need of antibiotic prophylaxis or suffering a contagious diseases were excluded.The study protocol has been presented in detail previously by Leppilahti et al. (2011)

| Oral rinse sampling and elastase activity test
Oral rinse samples for PMN elastase activity testing were collected before clinical examination.By means of a disposable plastic pipette, 1 mL of tap water was placed into the patient's mouth, and after rinsing for 1 min the rinse was collected into a tube (Nieminen et al., 1993;Uitto et al., 1996).Testing for PMN elastase activity was performed immediately after the oral rinse sample collection.The rest of the sample was immediately frozen for further analyses (Uitto et al., 1996).

| MMP-8 analyses
Quantitative aMMP-8 (IFMA) and TIMP-1 (ELISA) measurements were performed from the same oral rinse sample as the PMN elastase activity test.The results of aMMP-8 IFMA and TIMP-1 measurements were reported previously together with PMN elastase activity test results (Leppilahti et al., 2011).aMMP-8 IFMA measurements followed the same protocol described by Hanemaaijer et al. (1997) (Hanemaaijer et al., 1997).Briefly, the monoclonal aMMP-8-specific antibodies 8708 and 8706 were used as a catching antibody and a tracer antibody, respectively.The tracer antibody was labeled using europium-chelate (Hemmilä et al., 1984).Samples were diluted in assay buffer and incubated for 1 h, followed by incubation for 1 h with tracer antibody.
An enhancement solution was added, and after 5 min fluorescence was measured using a 1234 Delfia Research Fluorometer (Wallac).
From the 214 patients originally included, 192 frozen oral rinse samples were available for further WB for this study with the same principles described previously (Hanemaaijer et al., 1997;Kiili et al., 2002;Lauhio et al., 1994).
The molecular forms of MMP-8 were detected by using a modified enhanced chemiluminescence (ECL) WB kit according to the protocol recommended by the manufacturer (GE Healthcare) as described earlier (Gürsoy et al., 2018).Briefly: the oral rinse samples were mixed with Laemmli's buffer without any reducing reagents and heated for 5 min, followed by protein separation with 11% sodium dodecyl sulfate (SDS)-polyacrylamide gels.After electrophoresis, the proteins were electrotransferred onto nitrocellulose membranes Protran (Whatman GmbH, Dassel, Germany).Nonspecific binding was blocked with 5% milk powder (Valio Ltd.) in TBS-T buffer (10 mM Tris-HCl, pH 7.5, containing 22 mM NaCl and 0.05% Triton-X) for 1 h.
The membranes were incubated with monoclonal primary antibodies anti-aMMP-8 as by IFMA methods overnight, and then with horseradish peroxidase-linked secondary antibody (GE Healthcare) for 1 h.
The membranes were washed four times in TBS-T buffer between each step for 15 min.The proteins were visualized using the ECL system and scanned and analyzed using the GS-700 Imaging Densitometer Scanner (Bio-Rad) and Bio-Rad Quantity One program.
The analyses utilized the same catching antibody (8708) of the IFMA method for the identification of different molecular forms and species of MMP-8 by scanning image analysis.We regarded each MMP-8 molecular form finding as positive if a band of different molecular weights was observable and densitometric units showed value > 0. Because we wanted to study the correlation between quantitative MMP-8 levels and different protein forms and species in WB of the same oral rinse sample, it was more reasonable to use the same monoclonal antibody (catching ab 8708) utilized in the quantitative measurement and in the WB.

| Statistical methods
WB analysis is a semiquantitative immunological analysis method, and direct quantitative comparison between samples that are not analyzed in the same batch of immunoblotting and electrophoresis, is not advisable.However, the quantitative measurement and comparison of protein weights between samples within the same batch is Figure 1 shows the immunoblot analysis of recombinant latent proMMP-8 activated by NaOCl (Saari et al., 1990), organomercurial 4-Aminophenylmercuric acetate (APMA) (Saari et al., 1990), treponemal chymotrypsin-like protease (dentilisin) (Nieminen et al., 2018) (Figure 1a), and representative analysis of oral rinse samples from healthy participants and periodontitis patients (Figure 1b).All these known proMMP-8 activators activated proMMP-8 to lower molecular weight forms (40-60 kDa) and related low molecular size (25-40 kDa) fragments.
However, if latent MMP-8 was activated with a protease (dentilisin), a more versatile set of different-sized species can be observed in the range from 40 to 60 kDa (indicated as aMMP-8) and from 25 to 35kDa (fragments) (Figure 1a, lanes 8-10) corresponding with the biological/in vivo oral rinse samples (Figure 1b, lanes 2-5) of periodontitis patients.

| Correlation patterns in principal component analysis
Cumulative frequency distributions (CFD) of different MMP-8 species (percentual relative quantity) in WB express that the data T A B L E 1 Patient characteristics in relation to activated matrix metalloproteinase-8 (aMMP-8) immunofluorometric (IFMA) levels (tertiles).were right-skewed (Figure 2).In the CFD, the frequencies of (relative quantities of) each MMP-8 species are expressed cumulatively.The steeper the curve is in CFD, the more infrequent/rare the studied MMP-8 species are in comparison to other MMP-8 species.
Fragmented and 150 kDa complex forms were the most abundant MMP-8 species while the levels of 90-100 kDa complexes, 70-80 kDa proforms, and 45 and 55 kDa activated forms could not be detected in many samples.For this reason, the relative quantities of different protein weights were transformed with van der Waerden rank-based normalization to conduct principal component analysis.
In the PC analysis, distinct correlation patterns between MMP-8 molecular weights were observed (Table 2).The four first components explained 89% of the whole variance and there was no significant overlap between the components.PC1 correlated with fragments (21, 25 kDa), PC2 with the activated forms (45 and 55 kDa), and PC3 with proforms (70-80 kDa) and complexes (90-100 kDa).PC4 correlated with the fragment 35 kDa.There were significant negative correlations between 21 and 25 kDa fragments and 150 complexes, and the negative correlation of 150 complexes was loaded on the PC1 (Figure 3).

| Multilevel modeling of PCs
In multilevel models (Table 3), age and the number of sites with PPD 4-5 mm were negatively correlated with PC1, while elastase activity and TIMP-1 levels were correlating positively with PC1.Smoking was associated negatively with PC2, whereas PC4 was associated positively with smoking.Age and the number of teeth were correlating negatively with PC4.Age correlated positively with PC3.

| DISCUSSION
We found clear correlation patterns between the different MMP-8 WB molecular forms in PC analysis.The first four PCs explained 89% of the whole variance.For example, the fragmented 21 and 25 kDa forms were correlating negatively with the >150 kDa complexes.This suggests that MMP-8 appears more in fragmented forms in certain samples and in other cases MMP-8 is compounded into large molecular size complexes.Fragmented forms correlated strongly with the quantitative IFMA aMMP-8 levels, which has also been observed earlier in other data sets utilizing independent polyclonal anti-MMP-8 antibodies (Buduneli et al., 2011;Gürsoy et al., 2018).In multilevel models, different PCs of MMP-8 protein species were associated with smoking, age, PMN elastase activity, the number of teeth, and the number of moderately deep periodontal pockets (4-5 mm).
Especially fragmented low molecular size MMP-8 species and PMN elastase activity are surrogates and biomarkers of periodontal disease activity and collagenolytic and proteolytic inflammatory burden (Buduneli et al., 2011;Kiili et al., 2002;Leppilahti et al., 2011;Romanelli et al., 1999;Uitto et al., 1990).To the best of our knowledge, this is the first study analyzing correlations between WB and quantitative MMP-8 measurement utilizing the same antibody that is also used in the aMMP-8 point-of-care test (Deng et al., 2021(Deng et al., , 2022;;Lähteenmäki et al., 2022;Sorsa et al., 2020), and further on correlation between the different protein species in WB and patient-related factors and clinical surrogates.
The difference in the molecular weights of PMN and mesenchymal MMP-8 was explained with glycosylation of PMN-derived MMP-8.Autoproteolytic fragmentation of MMP-8 has been described previously (Knäuper et al., 1993).Activated MMP-8 (64 kDa) was autofragmented into a 40 and 27 kDa molecular form after 24 h incubation at 37°temperature.The fragmented MMP-8 was not able to cleave collagen I but the 40 kDa form had a similar proteolytic activity profile with the intact activated 64 kDa MMP-8 (Knäuper et al., 1993).The 27 kDa fragment had lost proteolytic activity after fragmentation.The 35 kDa band observed in our study was close to 40 kDa and could represent the autofragmented form of MMP-8.The precise mechanism producing lighter 21 and 25 kDa forms is not well known.However, the smaller sizes of MMP-8 species found in dental plaque are regarded to be due to proteolytic fragmentation by bacterial proteases (Kiili et al., 2002;Romanelli et al., 1999;Sorsa et al., 1995).T. denticola and other pathogens involved in dysbiosis with proteases such as dentilisin and gingipains can activate pMMP-8 (Nieminen et al., 2018;Sorsa et al. 1992Sorsa et al. , 1995)).Romanelli et al.
(1999) found these processed forms (55-60 kDa) of MMP-8 from GCF, which was sampled by oral rinse method from periodontitis patients.They considered these smaller MMP-8 forms as biologically important and hypothesized, that these species may be efficiently activated or even superactivated species.Especially P. gingivalis and T. denticola could catalyze the superactivation of MMP-8 by their F I G U R E 3 Scatterplot illustrating the correlations between the relative quantity (%) of different matrix metalloproteinase-8 (MMP-8) species and the regression factor of the first component (PC1) in principal component (PC) analysis.
In the present study, eight different bands were defined indicating the molecular weights of >150, 90-100, 70-80, 55, 45, 35, 25, and 21 kDa according to the previous literature, complex forms, different PMN proforms, activated forms, and three different fragmented forms.In this study, there was a distinct broader band around 90-100 kDa indicating probably MMP-8 complexes or so-called preproforms as suggested by Määttä et al. (2006).It is possible that the 90-100 kDa band includes also MMP-8 bound with TIMP-1 (28 kDa) as discussed below (Määttä et al., 2006;Owen et al., 2004).Such multiple molecular forms have also been identified in urine from diabetic nephropathic patients (Lauhio et al., 2008).However, only a minority of patients expressed a detectable band around 90-100 kDa (Figure 2).
The exact content of heavy complexes including MMP-8 and the mechanism leading to complex formation in oral fluid samples is not well known either.Reasonable explanations for heavy complexes could be that TIMP-1 is combined with MMP-8 (Määttä et al., 2006).The molecular weight of TIMP-1 is 28 kDa and if added to the 64 kDa weight of activated MMP-8 it sums up to around 90 kDa.Owen et al. (2004) also showed that TIMP-1 can be bound to the cellular membrane of PMN with unknown receptors and membrane-bound TIMP-1 that can further bind to MMP-8 via their COOH-terminal hemopexin domains (Owen et al., 2004;Wang et al., 2019).The membrane-bound TIMP-1-MMP-8 complex had interesting counterintuitive functional properties to promote pericellular proteolysis.TIMP-1 was also shown to enable MMP-8 and MMP-9, anchoring into neutrophil extracellular traps (NETs), extracellular DNA fibers restricting bacterial invasion and with bactericidal effects, facilitating proteolysis (Wang et al., 2019).
These membrane-bound or NET-associated MMP-8-TIMP1 complexes could explain the heavy MMP-8 complexes in WB possibly also including other unknown components.TIMP-1 (28 kDa) and (85 kDa) molecular weights sum up to 113 kDa, which is near 110 kDa.
Activated protein species could be observed in WB around 45 and 55 kDa.In some cases, clearly, distinct bands cannot be observed, but diffuse signals are observed in the range from 40 to 60 kDa corresponding to the experimentally activated latent MMP-8 with dentilisin (Figure 1a, lanes 8-10).In some studies, mesenchymal MMP-8 (experimental samples from endothelial cells and fibroblasts) is observed around 40-50 kDa as described above (Hanemaaijer et al., 1997).However, it can be assumed that MMP-8 in oral rinse samples is mainly derived from PMN leukocytes, as the cell type represents around 90% of cells in the GCF (Delima & van Dyke, 2003).
So, it is more reasonable to assume that the 45 kDa aMMP-8 species are proteolytically cleaved PMN-derived MMP-8.However, in the limits of the methods used in this study, we cannot verify, whether MMP-8 is of mesenchymal or PMN origin.
The PMN elastase activity was associated with the relative quantity of fragmented forms of MMP-8 and with aMMP-8 levels measured by IFMA and reflected periodontal inflammatory burden.
PMN elastase is a neutrophil-derived enzyme and PMN elastase activity probably indicates enzymatic activity eventually promoting and enhancing MMP-8 fragmentation.PMN elastase does not directly cleave MMP-8 but is strongly correlating with cathepsin G and myeloperoxidase, which are directly related to proteolytic or oxidative activation and fragmentation of MMP-8 and may also explain the correlation with fragmentation (Buduneli et al., 2011;Hernández et al., 2010;Leppilahti et al., 2014;Owen et al., 2004;Wang et al., 2019).
An interesting new finding was the association between age and different molecular forms of MMP-8 in WB.The PC1, which was correlating positively with the fragmented forms but negatively with the complexes had a negative significant correlation with age in the multilevel model.It seems that in younger participants, the relative quantity of fragmented MMP-8 is higher, while in older subjects MMP-8 is compounded into the heavy complexes.In turn, PMN leukocyte molecular forms (PC3) correlated positively with participants' age (Table 4).There are several possible explanations for the observed associations.Age is a well-known confounding factor of periodontitis prevalence, extent, and severity.We have not performed additional post hoc analysis that would consider age and periodontitis as confounding factors, but it can be speculated if periodontitis prevalence and severity would have modifying effects on MMP-8 turnover and processing in oral fluids.Deeper periodontal pockets correlate with the dysbiosis-associated periodontal key pathogens, which express MMP-8 activating enzymes gingipains (P.gingivalis), trypsins, and dentilisin (T.denticola) as described above (DeCarlo et al., 1997;Nieminen et al., 2018;Sorsa et al., 1992Sorsa et al., , 1995)).
A negative correlation between age and TIMP-1 levels was reported with this same data set when the association of periodontal parameters and quantitative aMMP-8 levels was reported previously (Leppilahti et al., 2011).Significant age-dependent association was observed in the ratio of aMMP-8 (IFMA) and TIMP-1 (ELISA) measurements but not in aMMP-8 (IFMA) measures itself.Several recent clinical trials, studying the accuracy of the aMMP-8 test, have reported that age is a significant cofactor affecting the aMMP-8 test accuracy (Deng et al., 2021;Räisänen et al., 2021).
Migration of neutrophils toward dental biofilms and release of NETs on supragingival biofilm has been observed and it can be assumed that NETs exist also in oral rinse samples (Hirschfeld et al., 2015;Preshaw et al., 2017).On the other hand, a significant age-related decrease in reactive oxygen species (ROS) and NET formation was observed after induction with lipopolysaccharide and interleukin-8 (Hazeldine et al., 2014).It can be speculated if agerelated changes in neutrophil functions, ROS, and NET release can have effects on MMP-8 activation and fragmentation versus complex formation also taking into account NET-related TIMP-1-MMP-8 complex discussed above (Wang et al., 2019).
A strength of this study was that both quantitative aMMP-8 (IFMA) measurements and WB analysis were performed by utilizing the same monoclonal antibody.However, it should be kept in mind that the immunofluorometric measurements as IFMA and sandwich ELISA methods utilize two monoclonal antibodies, called primary/ catching and secondary/detection antibodies.In WB, only one antibody is used, and the same primary/catching antibody (8708) was utilized in this WB analysis as was used in the quantitative IFMA.
This monoclonal antibody correlates well with the independent polyclonal antibody (Buduneli et al., 2011;Hernández et al., 2010), and both have been utilized to assess different forms of periodontal disease (Chen et 2000;Gürsoy et al., 2018;Kiili et al., 2002;Sorsa et al., 1999).The primary antibody was used because it is highly specific to catch the target, while the secondary antibody was not designed to be used for detection solely.It is also worth to keep in mind that although the antibodies are same as utilized in the | 1029 commercially available aMMP-8 point-of-care test (Deng et al., 2021(Deng et al., , 2022;;Lähteenmäki et al., 2022;Sorsa et al., 2020), the oral rinsing sampling method was different (Leppilahti et al., 2011) and their correlations with periodontitis-related factors, not about absolute quantities of these species.
Among the weaknesses of this study is that periodontal diagnosis does not comply with the new classification; clinical attachment level was not recorded, and the disease stage and grade could not be assessed.
Florida probe device was used for PPD and BOP measurements.The mean BOP percentages were surprisingly low in the study population if the same patients had deep periodontal pockets, had no periodontal treatment recently, and were seeking oral examination and treatment.
Thus, it can be speculated if there eventually was a systematic error in BOP detection, even though BOP percentages were low, there was a positive correlation between BOP percentage and MMP-8 (IFMA) levels which is in line with previous studies (Leppilahti et al., 2011;Räisänen et al., 2019).In the previous study with this same data set (Leppilahti et al., 2011), we could differentiate patients with a strong periodontal inflammatory burden, defined by an inflammatory burden index combining the information of the number of deep periodontal pockets and BOP percent, from periodontally healthy participants.However, as reported in the current study, there was solely a tendency but not significant association between IFMA aMMP-8 tertiles and number of periodontal pockets with PPD 4-5 mm and ≥6 mm (Table 1), which differs from most previous studies focusing on correlation/association between aMMP-8 and periodontal clinical surrogates (Rautava et al., 2020;Sorsa et al., 2016).
Nevertheless, the main interest in this study was to focus on the correlation between different MMP-8 forms and species in WB and quantitative aMMP-8 measurement (IFMA).
We conclude that fragmented MMP-8 species assessed by monoclonal antibodies are correlating with quantitative aMMP-8 measurements, as also found previously with an independent polyclonal antibody (Gürsoy et al., 2018), and there is a negative correlation between fragment and complex formation.Different environmental and patientrelated factors (smoking, age, proteolytic/PMN elastase activity) may modify the protein turnover in oral fluids and may cause some variability in quantitative aMMP-8 measurements.An interesting new finding was that age is possibly related to fragment and complex formation, which may also further modify quantitative MMP-8 measurement.Whether age is just a periodontitis-related confounding factor or has some direct immunosenescence-related mechanism affecting MMP-8 fragmentation/ complex formation, it cannot be answered with the explorative design of this study and needs to be validated with further clinical studies focusing more specifically on age-related changes in oral fluid biomarker levels.
. Concisely, oral rinse sampling for the PMN elastase activity test and oral examination, comprised of measurements of pocket probing depths (PPD) and of bleeding on probing (BOP) by a Florida probe carried out by two calibrated general dentists, were performed for all included patients.Clinical periodontal diagnoses were assessed according to the 1999 classification.Background characteristics were recorded by questionnaires before clinical examinations.All patients gave an informed consent.Ethical committees of the Institute of Dentistry, University of Helsinki, and Helsinki University Central Hospital accepted the study protocol.
possible.The patient (sample) specific relative percentual quantities of different protein weights in WB were calculated and further Van der Waerden rank-based normalization was performed to calculate correlations between all samples.Principal component (PC) analysis(Soloman & Sawilowsky, 2009) was conducted to explore correlation patterns of different WB MMP-8 protein forms.The polychoric correlation method and rotated Varimax method were used to extract PCs.Sidak's correlations between PCs and quantitative MMP-8 IFMA levels were calculated by adjusting for age, gender, and smoking.Multilevel regression analysis was conducted to test associations between WB principal components and periodontal disease-related surrogate measures and patient-related factors.Stata 16.0 (StataCorp LP) was used for all statistical analyses.Patient characteristics Background characteristics of study participants in relation to oral rinse aMMP-8 IFMA levels are described in Table 1.aMMP-8 levels correlated significantly with increasing elastase activity, BOP measures, and TIMP-1 levels.3.1.1| Western immunoblot analysis of proMMP-8 activation in oral rinse samples Different protein forms >150 kDa complexes, complexes (or preproforms) in the range from 90 to 100 kDa, 70-80 kDa latent (PMN) proforms, two activated forms/bands in 45 and 55 kDa, and 35, 25, 21 kDa fragmented forms and their relative (%) semiquantitative measures were observed and defined in WB.
T A B L E 4 Multilevel models of the western blot principal components (PC).
In this study, we solved this methodological problem by calculating the sample-specific relative (%) quantities of different MMP-8 species and we conducted further statistical analyses with these relative quantities of different molecular weights.So, it is important to keep in mind that we are speaking about the percentual share of different MMP-8 species in oral rinse samples