Urine and serum biomarkers associated with diabetic nephropathy

Translation of description of equivalent WO 2010085879 A1

Urine and Serum Biomarkers Associated with Diabetic Nephropathy

RELATED APPLICATION

This application claims pπoπty to U S Provisional Application No 61/147 778 filed on January 28, 2009, the content of which is heieby incorporated by reference in its entirety

BACKGROUND OF THE INVENTION

Diabetic nephropathy (DN) is a progressive kidney disease associated with longstanding diabetes melhtus It causes abnormal fluid filtration and increased uπnaiy albumin excretion, eventually leading to kidney failuie

DN displays no symptoms in its early course As such it is difficult to detect the incipiency of this disease In fact, present diagnosis of DN depends on development of microalbuminuria, which occurs when kidney damage is already in place The lack of an early diagnostic test prevents effective treatment of early stage DN It is of great importance to identify reliable biomarkers useful in diagnosing early stage

DN

SUMMARY OF THE INVENTION

The present invention is based on unexpected discoveπes that a numbei of urine and serum proteins and their fragments, either alone or in combination, are differentially presented in DN patients as compared to DN fiee subjects These protein molecules aie theiefore useful markeis for diagnosing early stage DN

Accordingly one aspect of this invention features a method of diagnosing DN in a subject This method includes at least two steps (i) detei mining in a subject suspected of having DN a level of a biomarker and (ii) assessing whether the subject has DN based on the level of the biomaikei An inciease in the level of the biomai kei , as compaied to that in a DN-free subject, indicates that the subject has DN

The biomarker used in this diagnostic method is one of the four protein molecules listed below (l) a first urine protein molecule that is piecursor alpha 2 HS glycoprotein or a fragment thereof having at least ten amino acid residues, such as, mature alpha-2-HS-glycoprotein VVSLGSPSGEVSHPRKT (SEQ ID NO 1 ), or MGVVSLGSPSGEVSHPRKT (SEQ ID NO 2) (π) a second urine protein molecule that is alpha 1 antitrypsin or a fragment thereof having at least ten amino acid residues, such as KGKWERPFEVKDTEEEDF (SEQ ID NO 3) MIEQNTKSPLFMGKVVNPTQK (SEQ ID NO 4), EDPQGDAAQKTDTSHHDQDHPTFNKITPNLAE (SEQ ID NO 5), or EDPQGDAAQKTDTSHHDQDHPTFNKITPNLAEFA (SEQ ID NO 6), (in) a third urine protein molecule that is a fragment of alpha- 1 acid glycoprotein having at least ten amino acid iesidues such as GQEHFAHLLILRDTKTYMLAFDVNDEKNWGLS (SEQ ID NO 7) and

(iv) a serum piotem molecule that is osteopontin or a fragment thereof having at least ten ammo acid residues, such as YPDAVATWLNPDPSQKQNLLAPQNAVSSEETNDFKQETLPSK (SEQ ID NO 8) or KYPDAVATWLNPDPSQKQNLLAPQTLPSK (SEQ ID NO 9)

The diagnostic method described above can further include, aftei the assessing step a step of coπelating the biomarker level with the DN status (i e , whether it is at early or late stage) When the biomaiker is piotein molecules (i) oi (iv), an increase in its level ielative to that in a DN-free subject is indicative of late stage DN For a biomarker that is protein molecules (ii) or (in), its level indicates the DN status when compared with pre determined ieterence biomaiker levels representing early and late stage DN

In another aspect the present invention features a method for assessing efficacy of a DN tieatment in a subject (e g , a human patient or a laboiatory animal) This method includes determining in the subject pre treatment and post treatment levels of protein molecules (i) (π), (in) or (iv) and assessing efficacy of the treatment based on a change in the level of the biomarker after the tieatment If the post tieatment level of the biomarkei remains the same oi decreases as compared to the pre-treatment level of the biomarker, it indicates that the treatment is effective In yet another aspect, this invention features a method for determining a DN stage, including at least four steps: (i) obtaining a urine sample and optionally, a serum sample from a subject suspected of having diabetic nephropathy, (ii) determining in the sample(s) a level of one of the biomarkers listed in the preceding paragraph, (iii) calculating a disease score based on the level of the biomarker, and (iv) assessing the subject's diabetic nephropathy stage based on the disease score as compared to pre-determined cutoff values indicating different diabetic nephropathy stages. In this method, the calculating step can be performed by ridge regression analysis, factor analysis, discriminant function analysis, and logistic regression analysis.

The biomarker used in the just-described DN staging method is composed of at least two of the following five protein molecules: protein molecules (i)-(iv) listed above and protein molecule (v) that is urine osteopontin or its fragment described above. In one example, the biomarker is composed of all of the five protein molecules. In another example, it is composed of at least two of protein molecules (i)-(iii) and (v).

Alternatively, the biomarker is composed of at least two of the five protein molecules listed above and additionally, one or more clinical factors, e.g., age, gender, HbAIc, albumin/creatinine ratio (ACR), and glomerular filtration rate (GFR).

In still another aspect, the present invention provides a method for monitoring DN progress based on the level of any of the above-mentioned biomarkers. This method includes obtaining two urine samples and optionally, two serum samples, within a time span of 2 weeks to 12 months (e.g., 2-24 weeks or 3- 12 months) from a subject suspected of having DN, determining in the samples a level of one of the biomarkers, calculating disease scores based on the biomarker levels, and assessing DN progress in the subject based on the disease scores. The disease score for the later-obtained samples being greater than that for the earlier-obtained samples is indicative of DN exacerbation. The biomarkers mentioned above can also be used to assess efficacy of a DN treatment.

The treatment is effective if the post-treatment level of one of the biomarkers remains unchanged or decreases as compared to the pre-treatment level of the same biomarker.

The present invention further provides a kit for use in any of the methods described above. This kit includes two, three, or four antibodies with different antigen specificities. Each of these antibodies is capable of binding to one ot (i) alpha 2-HS-glycoprotein, (u) alpha- 1 antitrypsin, (in) alpha 1 acid glycoprotein, and (iv) osteopontin In one example, this kit contains only antibodies specific to antigens to be detected (e g , biomarkers associated with DN) for practice one of the methods disclosed herein Namely, it consists essentially of such 5 antibodies

Also within the scope of this invention is an isolated antibody specifically binding one of the following peptide

MGVVSLGSPSGEVSHPRKT (SEQ ID NO 2) KGKWERPFEVKDTEEEDF (SEQ ID NO 3) l o MIEQNTKSPLFMGKVVNPTQK (SEQ ID NO 4),

EDPQGDAAQKTDTSHHDQDHPTFNKITPNLAEFA (SEQ ID NO 6), GQEHFAHLLILRDTKTYMLADVNDEKNWGLS (SEQ ID NO 7), YPDAVATWLNPDPSQKQ NLLAPQNAVSSEETNDFKQETLPSK (SEQ ID NO 8), and 15 KYPDAVATWLNPDPSQKQNLLAPQTLPSK (SEQ ID NO 9)

The terms 'an isolated antibody ' used herein refers to an antibody substantially free from naturally associated molecules More specifically, a preparation containing the antibody is deemed as "an isolated antibody ' when the naturally associated molecules in the preparation constitute at most 20% by diy weight Punty can be measured by any appropriate method, e g , 20 column chromatography, polyaci ylamide gel electrophoiesis, and HPLC

Any of the antibodies described above can be used in manufacturing a kit useful in practicing any of the methods of this invention

The details of one or moie embodiments of the invention are set forth in the descπption below Other features or advantages of the present invention will be apparent fiom the 25 following diawings and detailed desci iption of several embodiments and also tiom the appended claims

BRIEF DESCRIPTION OF THE DRAWING The drawing is first described Fig 1 is a diagram showing boxplots for urine alpha 2-HS glycoprotein (uDN2, see panel A) mine alpha 1 antitrypsin (uDN5 see panel B), urine alpha 1 acid glycoprotein (uGR3, see panel C), and serum osteopontin (sDNO see panel D) in various groups of DN patients The upper and lower limits of the boxes mark the 25% and 75% values with the medians as the lines across the boxes The upper whisker marks the largest value below the upper fence which is the 75% value plus 1 5 interquartile range and the lower whisker marks the smallest value above the lower fence which is the 25% value minus 1 5 interquartile range

DETAILED DESCRIPTION OF THE INVENTION

DN is a kidney disorder associated with diabetes It has five progression phases Stage 1 characterized by diabetic melhtus with noimal GFR and normal albuminuria

(ACR < 30 mg/g)

Stage 2 characterized by glomerular hyperfiltiation (greater than 120 mL/minute/1 73m²) and renal enlaigement accompanying with normal GFR and normal albuminuπa(ACR < 30 mg/g) Stage 3 chaiacteπzed by microalbuminuria

Stage 4 characterized by overt albuminuiia and a piogiessive decline in GFR and Stage 5 characterized by a GFR of less than 15 mL/minute/1 73m' Commonly stages 1 3 are deemed as early stage and stages 4 and 5 are deemed as late stage

We have identified a number of biomaikei s associated with DN especially DN in different stages These biomarkers are composed of one or more of the following four proteins and their fragments eithei in urine or in seium (a) alpha 2 HS glycoprotein (GenBank accession no NP_001613 10 Jan 2010) (b) alpha 1 antitrypsin (GenBank accession no AAB59495 10-Jan-2010) (c) alpha 1 acid glycoprotein (GenBank accession no EAW87416, 10 Jan 2010) and (d) Osteopontin which includes two isoforms known as secreted phosphoprotein Ia (GenBank accession no NP_001035147, 17 Jan 2010) and secieted phosphoprotein Ib (GenBank accession no NP_000573 10- Jan 2010 )

The fragments of these four proteins have a minimum length of ten amino acids and preferably a maximum length of 190 to 410 amino acids For example fragments of proteins (a), (b), (c), and (d) can contain up to 357, 408, 191 , and 290 amino acid residues, respectively. We have also found that biomarkers composed of one or more of the above mentioned proteins/fragments, and one or more clinical factors (e.g., age, gender, HbAIc, ACR, and GFR) are also associated with DN in different stages. Accordingly, one aspect of the present invention relates to a DN diagnostic method using any of the biomarkers described above. To practice this method, a urine sample and, when necessary, a serum sample, is collected from a subject suspected of having DN and the urine and serum levels of one or more of the four proteins listed above or their fragments can be determined via routine methods, e.g., mass spectrometry and immune analysis. If applicable, the clinical factors are determined by route methods.

When a biomarker contains a single protein molecule, its level in a subject can be compared with a reference point to determine whether that subject has DN. The reference point, representing the level of the same biomarker in a DN-free subject, can be determined based on the representative levels of the biomarker in groups of DN patients and DN-free subjects. For example, it can be the middle point between the mean levels of these two groups. A biomarker level higher than the reference point is indicative of DN.

When a biomarker contains at least two protein molecules and optionally, at least one clinical factor, the levels of the protein molecules and the value(s) of the clinical factor(s) can be subjected to a suitable analysis to generate a disease score (e.g., represented by a numeric number) that characterizes the level of the biomarkers. Examples of the analysis include, but are not limited to, discriminate function analysis, logistic regression analysis, ridge regression analysis, principal component analysis, factor analysis, and generalized linear model. The disease score is then compared with a reference point representing the level of the same biomarker in DN-free subjects. The reference point can be determined by conventional methods. For example, it can be a score obtained by analyzing the mean levels of the protein molecules and when necessary, the mean value(s) of the clinical factor(s) in DN-free subjects with the same analysis. The disease score being higher than the reference point is indicative of DN presence.

Another aspect of this invention relates to a method for determining a DN stage based on any of the biomarkers descπbed above To practice this method, a biomarker level of a DN patient, preferably represented by a disease score, is compared with a set of pre determined cutoff values that distinguish different DN stages to deteimine the subject s DN stage The cutoff values can be determined by analyzing the representative levels of the same biomarker in different staged DN patients via the same analysis

Descπbed below is an exemplary procedure for determining the aforementioned cutoff values based on a biomarker associated with DN in different stages

(1) assigning DN patients to different groups accoiding to their disease conditions (e g , DN stages and risk factors), (2) determining in each patient group the levels/values of the protein molecules and clinical factors constituting the biomarker

(4) subjecting the protein levels and clinical factor values to a suitable analysis to establish a model (e g , foi inula) for calculating a disease score, and

(6) determining a cutoff value for each disease stage based on a disease score (e g , mean value) representing each patient gioup as well as other relevant factors such as sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV)

Any of the models thus generated can be assessed for its diagnosis value by a receiver-operating characteristic (ROC) analysis to cieate a ROC curve An optimal multi variable model piovides a laige Area under Cuive (AUC) in the ROC analysis See the models described in Examples 1 3 below

In still another aspect this invention ielates to a method of monitoiing nephropathy progress in a subject based on any of the biomarkeis described above More specifically, two mine samples and/or seium samples from a subject can be obtained within a suitable time span (e g , 2 weeks to 12 months) and examined to determine the levels of one of the biomarkers Disease scores are then determined as descπbed above If the disease score repiesenting the biomarker level in the later obtained sample(s) is lower than that in the earlier-obtained sample(s), it indicates DN exacerbation in the subject

The monitoring method can be applied to a human subject suffering from or at risk for DN When the human subject is at risk for or in early stage DN, the level of the biomarkei can be examined once every 6 to 12 months to monitor DN progress When the human sub_ject is already in late stage DN it is preferred that the biomarker level be examined once every 3 to 6 months

The monitoring method descπbed above is also applicable to laboratory animals, following routine procedures to study DN The term 'a laboratory animal ' used herein refers to a vertebiate animal commonly used in animal testing, e g , mouse, rat, rabbit, cat dog pig, and non human primate Preferably, a laboratory animal is examined to determine the biomarker level once every 2 to 24 weeks

Any of the biomarkers can also be used to assess efficacy of a DN treatment in a subject in need (i e a human DN patient or a laboratoiy animal bearing DN) In this method disease scores representing levels of one of the biomarkers descπbed above aie determined before during, and after the treatment If the disease scores remain the same or decline over the course of the treatment, it indicates that the treatment is effective

Also disclosed herein is a kit useful in piacticing any of the above described methods This kit contains two, three, or four antibodies with different antigen specificities Each of these antibodies is capable of binding to one of (i) alpha 2 HS glycoprotein, (ii) alpha- 1 antitrypsin, (in) alpha- 1 acid glycoprotein or (iv) osteopontin The antibodies specific to proteins (i) (π) (in), and (iv) can bind to their fragments MGVVSLGSPSGEVSHPRKT (SEQ ID NO 2), KGKWERPFEVKDTEEEDF (SEQ ID NO 3), MIEQNTKSPLFMGKVVNPTQK (SEQ ID NO 4), EDPQGDAAQKTDTSHHDQDHPTFNKITPNLAEFA (SEQ ID NO 6), GQEHFAHLLILRDTKTYMLADVNDEKNWGLS (SEQ ID NO 7), YPDAVATWLNPDPSQKQNLLAPQNAVSSEETNDFKQETLPSK (SEQ ID NO 8), and KYPDAVATWLNPDPSQKQNLLAPQTLPSK (SEQ ID NO 9) i e , specific to any antibody epitopes contained in these fragments In one example, this kit contains only antibodies specific to antigens to be detected (e g , protein molecules associated with DN) for practice one of the methods disclosed herein Namely, the kit consists essentially of such antibodies

The kit descπbed above can include two different antibodies (i e , a coating antibody and a detecting antibody) that bind to the same antigen Typically, the detecting antibody is conjugated with a molecule which emits a detectable signal eithei on its own or via binding to another agent The term "antibody' used herein refers to a whole immunoglobulin or a fragment thereof, such as Fab or F(ab )•> that retains antigen-binding activity It can be naturally occurring or genetically engineeied (e g , single chain antibody chimeiic antibody, or humanized antibody) The antibodies included in the kit ot this invention can be obtained from commercial vendors Alternatively, they can be prepared by conventional methods See, for example, Harlow and Lane, (1988) Antibodies A Laboratory Manual, Cold Spring Harbor Laboratory, New York To produce antibodies against a particular biomarker as listed above, the marker, optionally coupled to a carrier protein (e g KLH) can be mixed with an adjuvant, and injected into a host animal Antibodies produced in the animal can then be purified by affinity chromatography Commonly employed host animals include rabbits, mice, guinea pigs, and rats Various adjuvants that can be used to inciease the immunological response depend on the host species and include Fieund s adjuvant (complete and incomplete), mineral gels such as aluminum hydroxide, CpG, surface active substances such as lysolecithin, pluronic polyols polyamons peptides, oil emulsions keyhole limpet hemocyanm, and dinitrophenol Useful human adjuvants include BCG (bacille Calmette-Gueπn) and Corynebacteiium parvum Polyclonal antibodies, i e heterogeneous populations of antibody molecules, are present in the sera of the immunized animal

Monoclonal antibodies i e , homogeneous populations of antibody molecules, can be prepared using standard hybiidoma technology (see, for example, Kohler et al (1975) Nature 256, 495, Kohler et al ( 1976) fcui J Immunol 6, 51 1 Kohlei et al ( 1976) Em J Immunol 6, 292 and Hammerling et al ( 1981 ) Monoclonal Antibodies and T Cell Hybπdomas, Elsevier N Y ) In particular, monoclonal antibodies can be obtained by any technique that provides for the production of antibody molecules by continuous cell lines in culture such as described in Kohler et al ( 1975) Nature 256, 495 and U S Patent No 4,376 1 10 the human B cell hybπdoma technique (Kosboi et al ( 1983) Immunol Today 4, 72 Cole et al ( 1983) Proc Natl Acad Sci USA 80, 2026, and the EBV hybπdoma technique (Cole et al (1983) Monoclonal Antibodies and Cancer Therapy, Alan R Liss Inc , pp 77 96) Such antibodies can be of any immunoglobulin class including IgG IgM IgE IgA, IgD and any subclass theieof The hybridoma producing the monoclonal antibodies of the invention may be cultivated in vitro or in vivo. The ability to produce high titers of monoclonal antibodies in vivo makes it a particularly useful method of production.

Moreover, antibody fragments can be generated by known techniques. For example, such fragments include, but are not limited to, F(ab')₂ fragments that can be produced by pepsin digestion of an antibody molecule, and Fab fragments that can be generated by reducing the disulfide bridges of F(ab')2 fragments.

Without further elaboration, it is believed that one skilled in the art can, based on the above description, utilize the present invention to its fullest extent. The following specific embodiments are, therefore, to be construed as merely illustrative, and not limitative of the remainder of the disclosure in any way whatsoever. All publications cited herein are incorporated by reference.

Example 1 : Diagnosing DN based on urine alpha-2-HS-glycoprotein, urine alpha-1 antitrypsin, urine alpha-1 acid glycoprotein, or serum osteopontin

Material and Methods (i) Subjects

83 diabetic mellitus patients (designated "DM subjects"), and 82 DN patients (designated "'DN subjects") were recruited at the Tri-General Military Hospital in Taipei, Taiwan, following the standards set forth by the American Diabetic Association and also described below:

DM suffering from diabetic melhtus but fiee of DN (see the standards described below) DN suffering from diabetic melhtus and secreting urinary protein at a level greater than 1 g per day, having DN as proven by biopsy, or having uremia

All of the subjects were assigned into a training group and a testing group at a ratio of 7 3

(ii) Sample collection and processing

First morning void uπnaiy samples and seium samples weie collected from each of the subjects mentioned above Peptides contained in the urine samples weie examined by urinary matπx assisted laser desorption/ ionization time-of flight mass spectrometry (MALDI-TOF-MS) and by isobaric tags for relative and absolute quantification (iTRAQ)

Protein molecules including alpha 2 HS glycoprotein (DN2) alpha 1 antitrypsin (DN5), osteopontin (DNO), and alpha- 1 acid glycoprotein (GR3), were examined to determine their concentrations in both the urine and serum samples by ELISA Briefly, uπne samples were mixed with protease inhibitors and diluted at 1 100 with a dilution buffer and the serum samples were diluted at 1 10 The diluted samples weie placed in ELISA plates in triplicates The levels of DNO DN2, DN5 and GR3 concentiations weie measuied via the standard sandwich ELISA method

A 5-parameter standard curve was used for concentration calculation Only standards and samples with %CV of less than 15 weie included, those not meeting criteria were iepeated The protein levels in the uπne samples were noimahzed against the creatinine levels in the same uune samples, which weie measuied with the Quantichiom Cieatinine Assay (BioAssay Systems, (Hayward) California, USA) (in) Statistical Analysis

The data indicating the urine and seium piotein concentrations of each examined protein was statistically analyzed and performed as repiesented by auROC fiom 0 44 0 87 in their independent ability to distinguish DN subjects from DM subjects

For each subject, correlation between values was detei mined by Spearman or Pearson analysis depending on results of test for normality Gioup mean or median comparisons were made with the Student T-test or the Nonparametπc Mann- Whitney Test as appropriate Statistical significance was obtained when p<0 05 Statistics were piesented eithei as mean + standard error of mean (SEM) or as median with [25%, 75%]

Results

(i) Patient Characteristics

Tables 1 and 2 below show the characteπstics of patients in the training group and testing group and those in DM, and DN groups

Table 1. Characteristics of Patients in Training and Testing Groups

Training Testing

P value (n=118) (n=47)

Age, mean(SD) 59 94 ( 9 37 ) 60 28 (948) 08362 Female, n (%) 83 (70) 27 (57) 016

MDRD_S_GFR, mean(SD) 86 56 ( 33 1 1 ) 83 05 ( 4396 ) 05785 ACR(ug/mg), mean(SD) 737 82 ( 1465 47 ) 1084 18 ( 203098 ) 02239 Urine TP/Cr(mg/mg), mean(SD) 1 01 ( 2 01 ) K 178) 09963 Seium Cieatinine (mg/dL), mean(SD) 1 02 ( 0 87 ) 1 34 (144) 00903 HbA Ic (%), mean(SD) 8 49 ( 1 5 ) 8 29 (219) 05356

Markers (creatinine-adjusted), mean (SD)

1452 71 1488 77 uDNO(ng/mg) 0 8687

( 1416 7 ) ( 1222 2 )

40 65 38 35 sDN0(ng/ml) 0 6926

( 34 52 ) ( 34 13 )

4225 77 5999 64 uDN2(ng/mg) 0 2983 ( 9279 63 ) ( 10305 95 )

15951 12 45479 82 uDN5(ng/mg) 0 3228 ( 94956 78 ) ( 199827 84 )

32823 47 42709 23 uGR3(ng/mg) 0 5333 ( 62290 96 ) ( 103787 54 )

Table 2. Characteristics of Patients in DM and DN Groups Training (n=l 18) Testing (n=47)

DM (n=61) DN (n=57) P value DM(n=22) DN(n=25) P value

5711 6296 5909 6132

Age, mean(SD) 00006 04230 (805) (98) (882) (1009)

Female, n (%) 43 (70) 40 (70) 100 12(55) 15(60) 093

11121 6018 1156 5441

MDRD SJ3FR, mean(SD) <0001 <0001

( 1575) ( 2559 ) ( 3366 ) ( 2979 )

1135 151526 963 202978

ACR(ug/mg), mean(SD) <0001 00004 (681) ( 181572) (561 ) (243231 )

Urine TP/Cr(mg/mg), 017 19 017 17 mean(SD) <0001 00019 (051) (256) (032) (218)

Serum Creatinine 066 142

00019 (mg/dL), mean(SD) (012) (112) <oooι 067 192 (015) (179)

834 87 837 822

HbAIc (%), mean(SD) 02311 08238 (148) (153) (161) (266)

Markers (creatinine-adjusted), mean (SD)

142218 136677 176954 151644 uDNO(ng/mg) 08083 05953 ( 110546) ( 134792) ( 126015) ( 19457)

2903 4617 262 6452 sDNO(ng/ml) 00026 00010

( 1932) (3732) ( 1153 ) ( 5047 )

196847 808487 96879 734869 uDN2(ng/mg) 00013 00074 (421858 ) ( 1310168) ( 114447 )( 1086595)

4003686 7180269

39024 33621 uDN5(ng/mg) (14718658 00467 (26486327 01899 ( 132763) ( 56808 )

) )

6747092

357606 244777 716931 uGR3(ng/mg) (10520828 <0001 00003 ( 135628) ( 274238 ) ( 8299686 ) )

Statistically significant differences in GFR, ACR piotein, and serum creatinine levels were observed in the DN subjects versus in the DM subjects There was no diffeience in gender distribution among the groups (ii) Piotein molecules associated with DN Via uπne proteomic analysis, the peptides listed in Table 3 below were found to be differentially presented in uπne samples from the DM subjects and DN subjects

Table 3. Differentially Presented Urine/Serum Peptides and Proteins in Which They are Located

<img file="WO2010085879A1_D0001.tif" />

Via ELISA analysis, three urine protein molecules i e , uDN2, uGR3, and uDN5, and one serum protein molecule, i e , sDNO were found to be associated with DN See Fig 1 , panels A D and Table 2 above More specifically the levels of uDN2 uDN5, uGR3, and sDNO were found to be elevated in DN subjects as compared with DMs (free of DN) indicating that they are reliable markers for DN Furthei, the levels of uDN5 and uGR3 in DN subjects exhibiting macioalbumiuiia (ACR &gt; 300 mg/g) weie highei than those in DN subjects exhibiting microalbumiuiia (ACR 30 mg/g to 300 mg/g) Macioalbumiuπa is an indicator of late stage DN and micioalbumiuπa indicates early stage DN

Example 2 Staging DN based on a combination of uDN2, uDN5, uGR3 uDNO, and sDNO Two-protein model

The combined levels of two of uDN2, uDN5, uGR3, uDNO, and sDNO in DM subjects, and DN subjects were subjected to discriminant function analysis logistic regiession analysis, and ridge iegression analysis The results from this study indicate that any combination of two of the five proteins or then fragments can be used as reliable markers for determining DN stages

Shown below is an exemplary two-protein model, i e , uDN5 and uGR3, including equations for calculating disease scores based on the combined levels of these two protein molecules. Also shown below are tables (i.e., Tables 4- 9) listing cutoff values, sensitivities, specificities, positive predictive values (PPV) and negative predictive values (NPV), and area under the ROC curve (AUROC) for this two-protein model.

Discriminant function analysis:

Disease Score = 0.3303 x log₂[uDN5](ng/mg) + 0.2732 x log₂[uGR3](ng/mg) +5

Table 4. Cutoff Values Representing DN Early and Late Stages Indicated by Urine Albumin Levels

Training set (n=118) Testing set (n=47)

DM , Micro DM , Micro albuminuria albuminuria

DM vs. DN vs. DM vs. DN vs. Macro Macro albuminuria albuminuria

Cut-off 1 1.227 1 1.691 1 1.227 1 1.691

Sensitivity (%) 93 93 96 100

Specificity (%) 90 90 77 83

PPV (%) 90 83 83 78

NPV (%) 93 96 94 100

AUROC 0.95 0.96 0.98 0.96

Table 5. Cutoff Values Representing DN Stages 1-5

Training set (n=118) 1 esting set (n=47)

DN -Stage 1 vs 2 5 1 2 vs 3-5 1-3 vs 4 5 1 4 vs 5 1 vs 2-5 1 -2 vs 3 5 1-3 vs 4-5 1 4 vs 5

Cut off 1 1 066 1 1 227 1 1 691 14 017 1 1 066 1 1 227 1 1 691 14 017

Sensitivity (%) 75 93 93 75 84 96 100 100

Specificity (%) 89 90 90 90 75 77 83 80

PPV (%) 92 90 83 21 87 83 78 18

NPV (%) 69 93 96 99 71 94 100 100

AUROC 0 86 0 95 0 96 0 95 0 9 0 98 0 96 0 91

Logistic regression analysis

Disease Scoie = exp (Logit value ) / ( 1 + exp (Logit value ) ), in which Logit_value = -12 5332 + 0 7197 x log₂[uDN51(ng/mg) +0 4941 x log₂[uGR3](ng/mg)

Table 6. Cutoff Values Representing DN Early and Late Stages Indicated by Urine Albumin Levels

Training set (n=l 18) Testing set (n=47)

DM , Micro

DM , Micro albuminuria albuminuria

DM vs. DN DM vs. DN vs. vs. Macro Macro albuminuria albuminuria

Cut off 0 445 0 676 0 445 0 676

Sensitivity (%) 93 93 100 100

Specificity (%) 90 90 82 83

PPV (%) 90 83 86 78

NPV (%) 93 96 100 100

AUROC 0 95 0 96 0 98 0 97 Table 7. Cutoff Values Representing DN Stages 1-5

Training set (n=118) Testing set (n=47)

DN -Stage 1 vs 2 5 1 2 vs 3 5 1 -3 vs 4 5 1 4 vs 5 1 vs 2 5 1 2 vb 3 5 1 3 vs 4 5 1 4 vs 5

Cut off 0 383 0 445 0 676 0 996 0 383 0 445 0 676 0 996

Sensitivity (%) 75 93 93 75 84 100 100 50

Specificity (%) 89 90 90 90 75 82 83 80

PPV (%) 92 90 83 21 87 86 78 10

NPV (%) 69 93 96 99 71 100 100 97

AUROC 0 86 0 95 0 96 0 95 0 9 0 98 0 97 0 88

Ridge iegression analysis

Disease Score = 1 7697+ 0 1520 x Iog-&gt;[uDN5](ng/mg) +0 2254 x log₂[uGR3](ng/mg)

Table 8. Cutoff Values Representing DN Early and Late Stages Indicated by Urine Albumin Levels

Training set (n=118) Testing set (n=47)

DM , Micro DM , Micro albuminuria albuminuria

DM vs. DN vs. DM vs. DN vs.

Macro Macro albuminuria albuminuria

Cut off 2 254 2 606 2 254 2 606

Sensitivity (%) 93 93 100 94

Specificity (%) 90 90 77 79

PPV (%) 90 83 83 74

NPV (%) 93 96 100 96

AUROC 0 94 0 96 0 98 0 96 Fable 9. Cutoff Values Representing DN Stages 1-5

Training set (n=118) Testing set (n=47)

DN-Stage 1 vs 2 5 1 2 vs 3 5 1-3 vs 4-5 1-4 vs 5 1 vs 2 5 1 -2 vs 3 5 1 3 vs 4 5 1 4 vs 5

Cut off 2 185 2 254 2 606 4 016 2 185 2 254 2 606 4 016

Sensitivity (%) 75 93 93 75 84 100 94 100

Specificity (%) 89 90 90 90 75 77 79 84

PPV (%) 92 90 83 21 87 83 74 22

NPV (%) 69 93 96 99 71 100 96 100

AUROC 0 86 0 94 0 96 0 95 0 89 0 98 0 96 0 91

Three-protein model

The combined levels of thiee of uDN2, uDN5, uGR3 uDNO and sDNO in DM subjects and DN subjects were subjected to disci iminant function analysis, logistic regression analysis, factor analysis, and πdge iegression analysis The iesults indicate that any three protein combination can be used as a reliable marker for DN staging

Shown below is an exemplary three-piotein model i e , uDN2, uDN5 and uGR3, including equations for calculating disease scores based on the combined levels of these three protein molecules Also shown below aie tables (i e , Tables 10 17) listing cutoff values sensitivities, specificities, PPV, NPV, and AUROC for this three piotein model

Disci iminant function analysis

Disease Score = 0 3340 x log₂[uDN5](ng/mg) 0 0142 x log₂[uDN21(ng/mg) + 0 2784 x log₂[uGR3](ng/mg) + 5

Table 10. Cutoff Values Representing DN Early and Late Stages Indicated by Urine Albumin Levels

Training set (n=118) Testing set (n=47)

DM , Micro DM , Micro albuminuria albuminuria

DM vs. DN vs. DM vs. DN vs. Macro Macro albuminuria albuminuria

Cut-off 1 1.190 1 1 663 I l 190 1 1 663

Sensitivity (%) 93 93 96 100

Specificity (%) 90 90 77 83

PPV (%) 90 83 83 78

NPV (%) 93 96 94 100

AUROC 0.95 0.96 0 98 0.96

Table 11. Cutoff Values Representing DN Stages 1-5

Training set (n=118) Testing set (n=47)

DN Stage 1 vs 2-5 1-2 vs 3-5 1 -3 vs 4-5 1-4 vs 5 I vs 2 5 1-2 vs 3-5 1 -3 vs 4-5 1 -4 vs 5

Cut otf 1 1.064 1 1.190 1 1 663 13 986 1 1 064 1 1 190 1 1.663 13 986

Sensitivity (%) 75 93 93 75 84 96 100 100

Specificity (%) 89 90 90 90 75 77 83 82

PPV (%) 92 90 83 21 87 83 78 20

NPV (%) 69 93 96 99 71 94 100 100

AUROC 0 87 0 95 0 96 0 95 0 9 0.98 0 96 0.91

Factor analysis Disease Score = 0 9190 x log₂[uDN5](ng/mg)+ 0 6997 x log₂|uDN2](ng/mg) +0 9003 x log,[uGR3](ng/mg)

Table 12. Cutoff Values Representing DN Early and Late Stages Indicated by Urine Albumin Levels

Training set (n=118) Testing set (n=47)

DM , Micro DM , Micro albuminuria albuminuria

DM vs. DN vs. DM vs. DN vs. Macro Macro albuminuria albuminuria

Cut off 26 356 28 057 26 356 28 057

Sensitivity (%) 84 93 88 100

Specificity (%) 90 90 91 86

PPV (%) 89 83 92 82

NPV (%) 86 96 87 100

AUROC 0 93 0 95 0 99 0 97

Table 13. Cutoff Values Representing DN Stages 1-5

Training set (n=l 18) Testing set (π=47)

DN Stage 1 vs 2 5 I 2 vs 3 5 1 3 vs 4 5 1 4 vs 5 1 vs 2 5 1 2 vs 3 5 1 3 vs 4-5 1 -4 vs 5

Cut off 25 669 26 356 28 057 36 464 25 669 26 356 28 057 36 464

Sensitivity (%) 68 84 93 75 84 88 100 50

Specificity (%) 89 90 90 90 88 91 86 84

PPV (%) 91 89 83 21 93 92 82 12

NPV (%) 63 86 96 99 74 87 100 97

AUROC 0 83 0 93 0 95 0 95 0 91 0 99 0 97 0 86

Logistic regression analysis

Disease Score = exp (Logit_value ) / ( 1 + exp (Logit_value ) ) in which Logit_value = ] 1 2820 + 0 881 O x log₂[uDN5](ng/mg) 0 3478 x log₂[uDN2](ng/mg) +0 5576 x log-&gt;[uGR3](ng/mg)

Table 14. Cutoff Values Representing DN Early and Late Stages Indicated by Urine Albumin Levels

Training set (n=118) Testing set (n=47)

DM , Micro DM , Micro albuminuria albuminuria

DM vs. DN vs. DM vs. DN vs. Macro Macro albuminuria albuminuria

Cut-off 0 462 0 798 0 462 0 798

Sensitivity (%) 91 88 96 94

Specificity (%) 90 90 82 83

PPV (%) 90 82 86 77

NPV (%) 92 93 95 96

AUROC 0 95 0 96 0 97 0 95

Table 15. Cutoff Values Representing DN Stages 1-5

Training set (n=118) Testing set (n=47)

DN Stage 1 vs 2-5 1 -2 vs 3 5 3 vs 4 5 1 4 vs 5 1 vs 2 5 1 2 vs 3 5 I 3 vs 4 5 1 4 vs 5

Cut off 0 361 0 462 0 798 0 997 0 361 0 462 0 798 0 997

Sensitivity (%) 75 91 88 75 90 96 94 100

Specificity (%) 89 90 90 90 75 82 83 82

PPV (%) 92 90 82 21 88 86 77 20

NPV (%) 69 92 93 99 80 95 96 100

AUROC 0 88 0 95 0 96 0 95 0 89 0 97 0 95 0 93

Ridge regression analysis

Disease Scoie = -1 2900 + 0 1800 x log₂[uDN5](ng/mg) 0 1013 x log₂luDN2](ng/mg) + 0 2505 x log₂[uGR3](ng/mg)

Table 16. Cutoff Values Representing DN Early and Late Stages Indicated by Urine Albumin Levels

Training set (n=118) Testing set (n=47)

DM , Micro DM , Micro albuminuria albuminuria

DM vs. DN vs. DM vs. DN vs.

Macro Macro albuminuria albuminuria

Cut-off 2.122 2.831 2 122 2.831

Sensitivity (%) 95 85 100 94

Specificity (%) 90 90 68 86

PPV (%) 90 81 78 81

NPV (%) 95 92 100 96

AUROC 0.95 0.95 0.97 0.95

Table 17. Cutoff Values Representing DN Stages 1-5

Training set (n=118) Testing set (n=47)

DN Stage 1 vs. 2-5 1 -2 vs.3-5 1 -3 vs 4-5 1 -4 vs 5 1 vs 2-5 1 -2 vs 3-5 1 -3 vs 4-5 1 -4 vs 5

Cut otf 2.083 2.122 2 831 3.943 2 083 2.122 2.831 3.943

Sensitivity (%) 78 95 85 75 87 100 94 100

Specificity (%) 89 90 90 90 69 68 86 82

PPV (%) 92 90 81 21 84 78 81 20

NPV (%) 71 95 92 99 73 100 96 100

AUROC 0.88 0.95 0.95 0 95 0.89 0.97 0.95 0.93

Four- protein model

The combined levels of four of uDN2, uDN5, uGR3, uDNO, and sDNO in DM subjects and DN subjects were subjected to discriminant function analysis, logistic regression analysis, factor analysis, and ridge regression analysis. The results indicate that any combination of four of the five proteins or their fragments can be used as a reliable marker for determining DN stages. Shown below is an exemplary four-protein model, i.e., uDN2, uDN5, uGR3, and sDNO, including equations for calculating disease scores based on the combined levels of these four protein molecules. Also shown below are tables (i.e., Tables 18-25) listing cutoff values, sensitivities, specificities, PPVs, NPVs, and AUROC for this four-protein model.

Discriminant function analysis: Disease Score = 0.2972 x log₂[uDN5](ng/mg) +0.0159 x log₂[uDN2](ng/mg) + 0.2014 x log₂[uGR3](ng/mg) + 0.5688 xlog₂[sDNO](ng/ml) +5

Table 18. Cutoff Values Representing DN Early and Late Stages Indicated by Urine Albumin Levels

Training set (n=118) Testing set (n=47)

DM , Micro DM , Micro albuminuria albuminuria

DM vs. DN vs. DM vs. DN vs.

Macro Macro albuminuria albuminuria

Cut-off 12.945 13.520 12.945 13.520

Sensitivity (%) 88 95 96 100

Specificity (%) 90 90 82 86

PPV (%) 89 83 86 82

NPV (%) 89 97 95 100

AUROC 0.94 0.96 0.97 0.97

Table 19. Cutoff Values Representing DN Stages 1-5

Training set (n=118) Testing set (n=47) DN-Stages 1 vs 2 5 1 2 vs 3 5 1 3 vs 4 5 1 4 vs 5 1 vs 2 5 1 2 vs 3 5 1 3 vs 4 5 1 4 vs 5

Cut-off 12 887 12 945 13 520 15 560 12 887 12 945 13 520 15 560

Sensitivity (%) 73 88 95 100 81 96 100 100

Specificity (%) 89 90 90 90 81 82 86 82

PPV (%) 91 89 83 27 89 86 82 20

NPV (%) 67 89 97 100 68 95 100 100

AUROC 0 87 0 94 0 96 0 97 0 93 0 97 0 97 0 89

Factor analysis

Disease Score = 0 9132 x log₂[uDN5](ng/mg) +0 6950 x log₂[uDN2](ng/mg) + 0 9080 x log₂[uGR3](ng/mg) + 0 4549 xlog₂fsDNO](ng/ml)

Table 20. Cutoff Values Representing DN Early and Late Stages Indicated by Urine Albumin Levels

1 raining set (n=l 18) I esting set (n=47)

DM , Micro DM , Micro albuminuria „,, albuminuria

DM vs. DN DM vs. vs. vs. Macro albuminuria Macro albuminuria

Cut off 28 459 30 095 28 459 30 095

Sensitivity (⁰Io) 82 93 92 100

Specificity (%) 90 90 91 83

PPV (%) 89 83 92 78

NPV (%) 85 96 91 100

AUROC 0 93 0 96 0 9^c ) 0 98

Table 21. Cutoff Values Representing DN Stages 1-5

Training set (n=118) Testing set (n=47)

DN-Stage 1 vs 2 5 I 2 vs 3-5 1 -3 vs 4 5 1 -4 vs 5 1 vs 2 5 1 2 vs 3-5 1 -3 vs 4-5 1-4 vs 5

Cut-off 28 347 28 459 30 095 38 624 28 347 28 459 30 095 38 624

Sensitivity (%) 67 82 93 75 81 92 100 50

Specificity (%) 89 90 90 90 94 91 83 84

PPV (%) 91 89 83 21 96 92 78 12

NPV (%) 62 85 96 99 71 91 100 97

AUROC 0 84 0 93 0 96 0 95 0 92 0 99 0 98 0 86

Logistic regression analysis

Disease Score = exp (Logit_value ) / ( 1 + exp (Logit_value ) ), in which Logit value = 13 7529 + 0 9460 x log₂[uDN51(ng/mg) -0 31 10 x log₂[uDN2](ng/mg) + 0 4957 x log₂| uGR3](ng/mg) + 0 4787 xlog₂[sDNO](ng/ml) fable 22. Cutoff Values Representing DN Early and Late Stages Indicated by Urine Albumin Levels

Training set (n=118) Testing set (n=47)

DM , Micro DM , Micro albuminuria albuminuria

DM vs. DN vs. DM vs. DN vs.

Macro Macro albuminuria albuminuria

Cut off 0 423 0 804 0 423 0 804

Sensitivity (%) 91 88 96 100

Specificity (%) 90 90 77 86

PPV (%) 90 82 83 82

NPV (%) 92 93 94 100

AUROC 0 96 0 96 0 97 0 96 Table 23. Cutoff Values Representing DN States 1-5

Training set (n=l 18) Testing set (n=47)

DN -Stages 1 vs. 2-5 l -2 vs.3-5 l -3 vs.4-5 1 -4 vs. 5 1 vs. 2-5 l-2 vs.3-5 1-3 vs.4-5 1 -4 vs. 5

Cut-off 0.341 0.423 0.804 0.998 0.341 0.423 0.804 0.998

Sensitivity (%) 75 91 88 75 90 96 100 100

Specificity (%) 89 90 90 90 75 77 86 82

PPV (%) 92 90 82 21 88 83 82 20

NPV (%) 69 92 93 99 80 94 100 100

AUROC 0.89 0.96 0.96 0.96 0.91 0.97 0.96 0.9

Ridge regression analysis:

Disease Score = -1.7588 + 0.1729 x log₂|uDN5](ng/mg) -0.0971 x log₂[uDN2](ng/mg) + 0.2381 x log₂luGR3](ng/mg) + 0.1312 xlog₂[sDNO](ng/ml)

Table 24. Cutoff Values Representing DN Early and Late Stages Indicated by Urine Albumin Levels

Training set (n=118) Testing set (n=47)

DM , Micro

DM , Micro albuminuria albuminuria

DM vs. DN DM vs. DN vs. vs. Macro Macro albuminuria albuminuria

Cut-off 2.261 2.854 2.261 2.854

Sensitivity (%) 91 85 96 94

Specificity (%) 90 90 77 90

PPV (%) 90 81 83 85

NPV (%) 92 92 94 96

AUROC 0.95 0.95 0.97 0.95 Table 25. Cutoff Values Representing DN Stages 1-5

Training set (n=118) Testing set (n=47)

DN-Stage 1 vs 2-5 1 2 vs 3 5 1-3 vs 4-5 1 4 vs 5 1 vs 2 5 1 2 vs 3 5 1 3 vs 4 5 1 4 vs 5

Cut off 2 079 2 261 2 854 3 950 2 079 2 261 2 854 3 950

Sensitivity (%) 77 91 85 75 87 96 94 100

Specificity (%) 89 90 90 90 69 77 90 82

PPV (%) 92 90 81 21 84 83 85 20

NPV (%) 70 92 92 99 73 94 96 100

AUROC 0 89 0 95 0 95 0 95 0 89 0 97 0 95 093

Five protein model

The combined levels of uDN2 uDN5, uGR3 uDNO, and sDNO in DM subjects and DN subjects were subjected to discriminant function analysis, logistic iegiession analysis factor analysis, and ridge regiession analysis The iesults indicate that the combination of these five proteins or their fragments can be used as a reliable marker for determining DN stages

Shown below are equations for calculating disease scores based on the combined levels of these five piotein molecules, as well as tables (i e , Tables 26 33) listing cutoff values, sensitivities specificities NPVs, PPVs, and AUROC for this five piotein model Discriminant function analysis

Disease Scoie = 0 2780 x log-&gt;[uDN5](ng/mg)+ 0 0231 x log-&gt;[uDN2](ng/mg) +0 2236 x log₂|uGR3](ng/mg) +0 6043 x log2[sDNO](ng/ml) 0 1513 xlog2[uDNO](ng/mg)+5

Table 26. Cutoff Values Representing DN Early and Late Stages Indicated by Urine Albumin Levels

Training set (n=118) Testing set (n=47)

DM , Micro DM , Micro albuminuria albuminuria

DM vs. DN vs. DM vs. DN vs. Macro Macro albuminuria albuminuria

Cut off 1 1 818 12 164 1 1 818 12 164

Sensitivity (%) 86 98 96 100

Specificity (%) 90 90 86 86

PPV (%) 89 83 89 82

NPV (%) 87 99 95 100

AUROC 0 94 0 97 0 98 0 98

Table 27. Cutoff Values Representing DN States 1-5

Training set (n=118) Testing set (n=47)

DN Stages 1 vs 2-5 1 -2 vs 3-5 1 -3 vs 4-5 1 4 vs 5 1 vs 2-5 1-2 vs 3-5 1 3 vs 4-5 1 4 vs 5

Cut-off 1 1 766 1 1 818 12 164 14 432 1 1 766 1 1 818 12 164 14 432

Sensitivity (%) 73 86 98 100 81 96 100 100

Specificity (%) 89 90 90 90 88 86 86 82

PPV (%) 91 89 83 27 93 89 82 20

NPV (%) 67 87 99 100 70 95 100 100

AUROC 0 86 0 94 0 97 0 98 0 94 0 98 0 98 0 91 Factor analysis:

Disease Score = 0.91 17 x log₂[uDN5](ng/mg)+ 0 6949 x log₂[uDN2](ng/mg)+ 0.9095 x log₂[uGR3](ng/mg) + 0.4554 x log2[sDNO](ng/ml)+ 0.0384 xlog2[uDNO](ng/mg)

Table 28. Cutoff Values Representing DN Early and Late Stages Indicated by Urine Albumin Levels

Training set (n=118) Testing set (n=47)

DM , Micro DM , Micro albuminuria albuminuria

DM vs. DN vs. DM vs. DN vs.

Macro Macro albuminuria albuminuria

Cut off 29 475 30 541 29 475 30 541

Sensitivity (%) 81 93 88 100

Specificity (%) 90 90 91 83

PPV (%) 88 83 92 78

NPV (%) 83 96 87 100

AUROC 0 93 0 96 0 99 0 98

Table 29. Cutoff Values Representing DN Stages 1-5

Training set (n=l 18) Testing set (n=47)

DN Stages 1 vs 2-5 1 2 vs 3-5 1 3 vs 4-5 1-4 vs 5 1 vs 2-5 1 2 vs 3 5 1 -3 vs 4-5 1 -4 vs 5

Cut-off 28 740 29.475 30 541 39 042 28 740 29 475 30 541 39 042

Sensitivity (%) 67 81 93 75 81 88 100 50

Specificity (%) 89 90 90 90 94 91 83 84

PPV (%) 91 88 83 21 96 92 78 12

NPV (%) 62 83 96 99 71 87 100 97

AUROC 0 84 0 93 0 96 0 95 0 92 0 99 0 98 0 86 Logistic regression analysis

Disease Score = exp (Logit value ) / ( 1 + exp (Logit_value ) ), in which

Logit_value = 11 4318 + 0 8188 x log₂[uDN5](ng/mg) 0 5376 x log₂[uDN2](ng/mg) +07561 x log₂[uGR3](ng/mg) +0 3940 xlog₂[sDNO](ng/ml) -0 1741 xlog₂[uDNO](ng/mg)

Table 30. Cutoff Values Representing DN Early and Late Stages Indicated by Urine Albumin Levels

T raining set (n=118) Testing set (n=47)

DM , Micro DM , Micro albuminuria albuminuria

DM vs. DN vs. DM vs. DN vs.

Macro Macro albuminuria albuminuria

Cut off 0 436 0 780 0 436 0 780

Sensitivity (%) 91 93 96 100

Specificity (%) 90 90 77 86

PPV (%) 90 83 83 82

NPV (%) 92 96 94 100

AUROC 0 96 0 96 0 97 0 96

Table 31. Cutoff Values Representing DN States 1-5

Training set (n=118) Testing set (n=47)

DN Stages 1 vs 2-5 1 2 vs 3 5 1 3 vs 4 5 1 4 vs 5 1 vs 2 5 1 2 vs 3-5 1 3 vs 4 5 1 4 vs 5

Cut off 0 329 0 436 0 780 0 997 0 329 0 436 0 780 0 997

Sensitivity (%) 75 91 93 100 90 96 100 100

Specificity (%) 89 90 90 90 75 77 86 80

PPV (%) 92 90 83 27 88 83 82 18

NPV (%) 69 92 96 100 80 94 100 100

AUROC 0 89 0 96 0 96 0 96 0 91 0 97 0 96 0 91 Ridge iegiession analysis

Disease Score = 1 31 12 + 0 1648 x logi[uDN5](ng/mg) 0 0968 x log₂[uDN2](ng/mg) + 0 2468 x log₂[uGR3](ng/mg) + 0 1426 x log2[sDNO](ng/ml) -0 0552 xlog2[uDNO](ng/mg)

Table 32. Cutoff Values Representing DN Early and Late Stages Indicated by Urine Albumin Levels

Training set (n=118) 1 esting set (n=47)

DM , Micro DM , Micro albuminuria albuminuria

DM vs. DN vs. DM vs. DN vs.

Macro Macro albuminuria albuminuria

Cut-off 2 244 2 729 2 244 2 729

Sensitivity (%) 91 88 96 100

Specificity (%) 90 90 82 90

PPV (%) 90 82 86 86

NPV (%) 92 93 95 100

AUROC 0 95 0 95 0 98 0 97

Table 33. Cutoff Values Representing DN Stages 1-5

Training set (n=118) Testing set (n=47)

DN Stages 1 vs 2 5 1 2 vs 3 5 1 3 vs 4 5 1 4 vs 5 1 vs 2 5 1 2 vs 3-5 1 3 vs 4 5 1 -4 vs 5

Cut-off 2 043 2 244 2 729 3 913 2 043 2 244 2 729 3 913

Sensitivity (%) 77 91 88 100 87 96 100 100

Specificity (%) 89 90 90 90 69 82 90 80

PPV (%) 92 90 82 27 84 86 86 18

NPV (%) 70 92 93 100 73 95 100 100

AUROC 0 89 0 95 0 95 0 96 0 9 0 98 0 97 0 93 Example 3: Staging DN based on a combination of uDN2. uDN5, uGR3, and age

Shown below are equations for calculating disease scores determined by discriminant function analysis, factor analysis, logistic regression analysis, and ridge regression analysis, based on the level of a biomarker composed of three protein molecules, i.e., uDN2, uDN5, and uGR3, and one clinical factor, i.e., age. Also shown below are tables (i.e., Tables 34-41) listing cutoff values, sensitivities, specificities, PPVs, NPVs, and AUROC for this model.

Discriminant function analysis:

Disease Score = 0.3342 x log₂[uDN5](ng/mg) -0.0201 x log₂[uDN2](ng/mg) + 0.2826 x log₂[uGR3](ng/mg) + 0.0059 x Age(year) + 5

Table 34. Cutoff Values Representing DN Early and Late Stages Indicated by Urine Albumin Levels

Training set (n=118) Testing set (n=47)

DM , Micro DM , Micro albuminuria albuminuria

DM vs. DN vs. DM vs. DN vs. Macro Macro albuminuria albuminuria

Cut-off 1 1.515 12.088 1 1.515 12.088

Sensitivity (%) 93 93 100 100

Specificity (%) 90 90 77 79

PPV (%) 90 83 83 75

NPV (%) 93 96 100 100

AUROC 0.95 0.96 0.98 0.97

Table 35. Cutoff Values Representing DN Stages 1-5 Training set (n=118) Testing set (n=47)

DN -Stage 1 vs. 2-5 l -2 vs.3-5 1-3 vs.4-5 1 -4 vs. 5 1 vs. 2-5 l -2 vs.3-5 1-3 vs.4-5 1 -4 vs. 5

Cut-off 1 1.353 1 1.515 12.088 14 343 1 1.353 1 1.515 12.088 14.343

Sensitivity (%) 75 93 93 75 84 100 100 100

Specificity (%) 89 90 90 90 75 77 79 80

PPV (%) 92 90 83 21 87 83 75 18

NPV (%) 69 93 96 99 71 100 100 100

AUROC 0.87 0.95 0.96 0 95 0 9 0.98 0 97 0 9

Factor analysis:

Disease Score = 0.9184 x log₂| uDN5](ng/mg) +0.7006 x log₂[uDN2](ng/mg) + 0.9005 x log₂LuGR3 ](ng/mg) + 0.1863 x Age(year)

Table 36. Cutoff Values Representing DN Early and Late Stages Indicated by Urine Albumin Levels

Training set (n=118) Testing set (n=47)

DM , Micro

DM , Micro albuminuria albuminuria

DM vs. DN DM vs. DN vs. vs. Macro Macro albuminuria albuminuria

Cut-oft 38 341 40 075 38 341 40.075

Sensitivity (%) 82 85 96 100

Specificity (%) 90 90 86 83

PPV (%) 89 81 89 78

NPV (%) 85 92 95 100

AUROC 0.93 0 94 0 99 0.98

Table 37. Cutoff Values Representing DN Stages 1-5 rl raining set (n=118) Testing set (n=47)

DN-Stages 1 vs 2 5 1 2 vs 3-5 1-3 vs 4 5 1 4 vs 5 1 vs 2-5 1 -2 vs 3-51 3 vs 4-5 1 4 vs 5

Number of patients (%)

73 (62) 57 (48) 41 (35) 4 (3) 3 I (66) 25 (53) 18 (38) 2 (4)

Cut-off 38 341 38 341 40 075 48 538 38 341 38 341 40 075 48 538

Sensitivity (%) 66 82 85 50 81 96 100 50

Specificity (%) 89 90 90 90 88 86 83 89

PPV (%) 91 89 81 15 93 89 78 17

NPV (%) 62 85 92 98 70 95 100 98

AUROC 0 82 0 93 0 94 0 91 0 9 0 99 0 98 0 77

Logistic regiession analysis

Disease Score - exp (Logit_value ) / ( 1 + exp (Logit_value ) ), in which Logit value = 15 9748 + 0 8688 x log₂[uDN5](ng/mg) 0 4966 x log9[uDN2](ng/mg) + 0 6436 x log₂[uGR3](ng/mg) + 0 0879 x Age(yeai)

Table 38. Cutoff Values Representing DN Early and Late Stages Indicated by Urine Albumin Levels

Training set (n=118) Testing set (n=47)

DM , Micro DM , Micro albuminuria albuminuria

DM vs. DN vs. DM vs. DN vs. Macro Macro albuminuria albuminuria

Cut off 0 321 0 889 0 321 0 889

Sensitivity (%) 93 80 100 94

Specificity (%) 90 90 77 83

PPV (%) 90 80 83 77

NPV (%) 93 90 100 96

AUROC 0 96 0 95 0 97 0 95 Table 39. Cutoff Values Representing DN Stages 1-5

Training set (n=118) Testing set (n=47)

DN-Stages 1 vs 2 5 1 2 vs 3-5 1 3 vs 4 5 1 4 vs 5 1 vs 2 5 1 2 vs 3 5 1 3 vs 4 5 1 4 vs 5

Cut off 0 301 0 321 0 889 0 997 0 301 0 321 0 889 0 997

Sensitivity (%) 75 93 80 75 87 100 94 100

Specificity (%) 89 90 90 90 75 77 83 89

PPV (%) 92 90 80 21 87 83 77 29

NPV (%) 69 93 90 99 75 100 96 100

AUROC 0 89 0 96 0 95 0 92 0 88 0 97 0 95 0 91

Ridge regression analysis

Disease Scoie = 2 1690 + 0 1771 x log9[uDN5](ng/mg) 0 1074 x log₂|uDN2|(ng/mg) + 0 2474 x logi[uGR3](ng/mg) + 0 0168 x Age(year)

Table 40. Cutoff Values Representing DN Early and Late Stages Indicated by Urine Albumin Levels

Training set (n=l 18) Testing set (n=47)

DM , Micro

DM , Micro albuminuria albuminuria

DM vs. DN DM vs. DN vs. vs. Macro Macro albuminuria albuminuria

Cut off 2 139 2 880 2 139 2 880

Sensitivity (%) 93 85 100 89

Specificity (%) 90 90 73 83

PPV (%) 90 81 81 76

NPV (%) 93 92 100 92

AUROC 0 96 0 95 0 98 0 96 Table 41. Cutoff Values Representing DN Stages 1-5

Training set (n=118) Testing set (n=47)

DN-Stage 1 vs 2-5 1 2 vs 3 5 I 3 vs 4 5 1 4 vs 5 1 vs 2 5 1 2 vs 3 5 3 vs 4 5 1 -4 vs 5

Cut off 2 128 2 139 2 880 4051 2 128 2 139 2 880 4 051

Sensitivity (%) 75 93 85 75 84 100 89 100

Specificity (%) 89 90 90 90 69 73 83 89

PPV (%) 92 90 81 21 84 81 76 29

NPV (%) 69 93 92 99 69 100 92 100

AUROC 0 89 0 96 0 95 0 92 0 89 0 98 0 96 0 92

OTHER EMBODIMENTS

All of the features disclosed in this specification may be combined in any combination Each feature disclosed in this specification may be replaced by an alternative feature serving the same, equivalent, or similar purpose Thus, unless expressly stated otheiwise, each featuie disclosed is only an example of a generic series of equivalent or similar features

Fi oiii the above descπption one skilled in the ai t can easily ascertain the essential characteristics of the present invention, and without departing fiom the spirit and scope thei eof can make various changes and modifications of the invention to adapt it to various usages and conditions Thus other embodiments are also within the claims