Assay to detect and quantitate specific an antibodies for various redox forms of HMGB1

Abstract

Quantitation of specific antibodies for at least one redox form of High mobility group box 1 (HMGB1) contained in a biological sample. An in vitro method for assessing the state of progression of a disease or a disorder in which HMGB1 is involved. An in vitro method for the identification of predisposition, prognostic or diagnostic biomarkers of a disease or a disorder in which HMGB1 is involved. A kit to quantitate said specific antibodies for at least one redox form of HMGB1.

Claims

1. An in vitro method for quantitating specific antibodies for at least one redox form of High mobility group box 1 (HMGB1) contained in a biological sample obtained from a subject comprising: a) contacting said biological sample with two redox forms of HMGB1; and b) quantitating the specific antibodies for the two redox forms of HMGB1; wherein the redox forms of HMGB1 are selected from the group of fully reduced HMGB1 in which all three cysteine residues at positions 23, 45 and 106 have been completely reduced to thiols, disulfide-HMGB1 having a disulfide bridge between cysteine residues at positions 23 and 45 and a reduced cysteine residue at position 106, and oxidized-HMGB1 in which the three cysteine residues at positions 23, 45 and 106 have been oxidized to sulfonates.

2. The in vitro method according to claim 1, comprising before the step of contacting the biological sample with the two redox forms of HMGB1, a step of treating the biological sample by an acid solution to dissociate the immune complexes found in the sample, wherein in said method, the quantitated specific antibodies for the two redox forms of HMGB1 are total specific antibodies for the two redox forms of HMGB1.

3. The in vitro method according to claim 1, wherein the quantitated specific antibodies for the two redox forms of HMGB1 are their circulating fraction (circulating antibodies) and/or their immunologically complexed fraction.

4. The in vitro method according to claim 1, wherein said at least one redox form of HMGB1 has a mammalian origin.

5. The in vitro method according to claim 1, wherein said biological sample is blood, plasma, serum, saliva, peripheral blood mononuclear cells (PBMCs) or PBMC supernatant, CerebroSpinal Fluid (CSF) and/or wherein said biological sample is diluted from 1/800 to 1/2000.

6. The in vitro method according to claim 1, wherein the quantities of specific antibodies for the two redox forms of HMGB1 are determined by ELISA using the two redox forms of HMGB1 coated on solid supports.

7. The in vitro method according to claim 6, wherein said method comprises the steps of: coating a solid support with at least one redox form of HMGB1, redox form of HMGB1 is oxidizcd HMGB1; washing the solid support; blocking unbound sites with a saturation buffer; optionally, treating the biological sample by an acid solution to dissociate the immune complexes found in the sample; adding the biological sample to said solid support; optionally, adding a secondary antibody having binding capacity for the quantitated antibodies in complex with a redox form of HMGB1; quantitating the specific antibodies for at least one redox form of HMGB1.

8. The in vitro method according to claim 1: for quantitating specific antibodies for fully reduced HMGB1, wherein the method comprises a step of treating fully reduced HMGB1 to stabilize the reduced state of HMGB1, said treatment encompassing adding a reducing agent to fully reduced HMGB1; or for quantitating specific antibodies for oxidized-HMGB1, wherein the method comprises a step of treating disulfide-HMGB1 to obtain oxidized-HMGB1, said treatment encompassing adding a reactive oxygen species (ROS)[,] to disulfide-HMGB1.

9. The in vitro method according to claim 1, comprising: individually quantitating specific antibodies for fully reduced HMGB1, disulfide-HMGB1 and oxidized-HMGB1; individually quantitating specific antibodies for oxidized HMGB1 and specific antibodies for fully reduced HMGB1; or individually quantitating specific antibodies for oxidized HMGB1 and specific antibodies for disulfide HMGB1.

10. The in vitro method of claim 1, wherein the sample is contacted with fully reduced HMGB1, disulfide-HMGB1 and oxidized-HMGB1.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) FIG. 1. Redox state of HMGB1 and functions.

(2) FIG. 2. Principle of the assay.

(3) FIGS. 3A, 3B and 3C. Interference in the assay. The plate was coated with 3 μg/ml non-oxidizable HMGB1 in Dulbecco's Phosphate-Buffered Saline (DPBS) and HMGB1-specific antibodies concentrations were determined on serial dilutions (from 1/45 to 1/210) of a serum from a healthy donor (HD).

(4) FIG. 4. Influence of deoxycholic acid concentrations on interference. The plate was coated with 3 μg/mL non-oxidizable HMGB1 in

(5) DPBS or carbonate-bicarbonate buffer containing different concentrations of deoxycholic acid. HMGB1-specific antibodies concentrations were determined on serial dilutions (from 1/50 to 1/210) of a HD serum.

(6) FIG. 5. Effect of TWEEN® 20 on interference. Plates were coated with 3 μg/ml non-oxidizable HMGB1 in DPBS. Increasing concentrations of TWEEN® 20 were added to the washing buffer and the antibody dilution buffer. HMGB1-specific antibodies concentrations were determined on serial dilutions (from 1/50 to 1/210) of a HD serum.

(7) FIG. 6. Effect of serum dilution on interference. Plates were coated with 3 μg/ml non-oxidizable HMGB1 in DPBS. HMGB1-specific antibodies concentrations were determined on serial dilutions of different HD sera, from 1/3 to 1/6000 (A) or from 1/800 to 1/2000 (B).

(8) FIG. 7. Acidic treatment for the detection of total anti-HMGB1 antibodies in human samples. Human sera, either untreated (Unt) or treated (Tr) with Glycine 1.5M, were titrated for the presence of anti-HMGB1 IgG antibodies. Circulating residual anti-HMGB1 antibodies were detected at very low levels, while total anti-HMGB1 antibodies, including the complexed ones were detected at high levels following the acidic treatment. The plates were coated with 3 μg/ml non-oxidizable HMGB1 in DPBS.

(9) FIGS. 8A and 8B. Reproducibility. The serum from one HD was used as internal control in different experiments to test the reproducibility of the assay. The plates were coated with 3 μg/ml non-oxidizable HMGB1 in DPBS.

(10) FIG. 9. Impact of DTT on the specific detection of various redox forms of HMGB1. Different forms of HMGB1 were run on a 12% polyacrylamide gel in the presence or the absence of 1 mM DTT (A). Plates were coated with 3 μg/ml of the different redox forms of HMGB1, as provided by HMGBiotech. In order to keep the reduced form of HMGB1, 1 mM DTT was added to the coating, saturation and antibody dilution buffers (B) or only during coating and saturation steps (C). 100 mM H.sub.2O.sub.2 was added to the coating buffer in order to obtain the oxidized form of HMGB1. The specificity of two different murine anti-HMGB1 monoclonal antibodies was tested: DPH1.1 (B) and 1E6-E10 (C).

(11) FIG. 10. Ability of the assay to determine the specificity of antibodies towards the various redox forms of HMGB1. Plates were coated with 3 μg/ml of the different redox forms of HMGB1, as provided by HMGBiotech. 1 mM DTT was added to the coating and saturation buffers in order to obtain the all-thiol form of HMGB1. 100 mM H.sub.2O.sub.2 was added to the coating buffer in order to obtain the oxidized form of HMGB1. The specificity of the mouse monoclonal antibody anti-HMGB1 1E6-E10 was tested on the different forms of the protein.

(12) FIG. 11. Detection and quantification of antibodies to the mutant, non-oxidizable (Non-Ox) form of HMGB1, with or without DTT, in HD sera. Plates were coated with 3 μg/ml of the non-oxidizable form of HMGB1. 1 mM DTT was added to the coating and saturation buffers in order to verify the effect of DTT on the binding capacity of the antibodies. Titration of total antibodies recognizing the non-oxidizable form of HMGB1 was made in the serum of 9 HD (serum dilution 1/1500).

(13) FIG. 12. Detection and quantification of antibodies to the mutant, non-oxidizable (Non-Ox) form of HMGB1, with different concentrations of DTT, in HD sera. Plates were coated with 3 μg/ml of the non-oxidizable form of HMGB1. 0.5 mM, 1 mM or 5 mM DTT was added to the coating and, for selected samples, to the saturation buffers in order to verify the effect of DTT on the antibodies binding capacity. Titration of total antibodies recognizing the non-oxidizable form of HMGB1 was made in the serum of 3 HD (serum dilution 1/1000).

(14) FIG. 13. Detection and quantification of antibodies to the different forms of HMGB1 in HD sera. Plates were coated with 3 μg/ml of the different forms of HMGB1. 0.5 mM DTT was added to the coating buffer in order to obtain the all-thiol form of HMGB1. 100 mM H.sub.2O.sub.2 was added to the coating buffer in order to obtain the oxidized form of HMGB1. Titration of total specific antibodies for the different forms of HMGB1 was made in the serum of 5 HD (serum dilution 1/1000).

(15) FIG. 14. Serum levels of specific antibodies for the different forms of HMGB1 in healthy donors (HD) and HIV.sup.+ Patients. Plates were coated with 3 μg/ml of the different forms of HMGB1. 0.5 mM DTT was added to the coating buffer in order to obtain the all-thiol form of HMGB1. 100 mM H.sub.2O.sub.2 was added to the coating buffer in order to obtain the oxidized form of HMGB1. Titration of total specific antibodies for the different forms of HMGB1 was made in the serum of 29 HD and 68 HIV-infected patients (serum dilution 1/1000). Statistical significance for comparison of the 2 groups is shown: ns: not significant, **p<0.01, ****p<0.0001.6

(16) FIG. 15. Serum concentration of specific antibodies for the different forms of HMGB1 in patients with HIV-associated neurological disorders (HAND). Plates were coated with 3 μg/ml of the indicated forms of HMGB1. 0.5 mM DTT was added to the coating buffer in order to obtain the all-thiol form. 100 mM H.sub.2O.sub.2 was added to the coating buffer in order to obtain the oxidized form of HMGB1. Titration of total specific antibodies for the different forms of HMGB1 was made in the serum of 29 HD and 68 HIV-infected patients. Groups of patients were defined according to their clinical neurological status (group 1: stage 1, no HAND, n=22; group 2: stage 2, no HAND with single deficit, n=26; group 3: stage 3, ANI (Asymptomatic Neurological Impairment), n=12; group 4: stage 4, MND+HAD (Mild Neurological Disorders+HIV-Associated Dementia), n=8). Statistical significance is shown: **p<0.01, ***p<0.001, ****p<0.0001.

(17) FIG. 16. Comparison of levels of specific antibodies for the different forms of HMGB1 in healthy donors (HD) and HIV.sup.+ patients with neurological disorders. Plates were coated with 3 μg/ml of the indicated forms of HMGB1. 0.5 mM DTT was added to the coating buffer in order to obtain the all-thiol form. 100 mM H.sub.2O.sub.2 was added to the coating buffer in order to obtain the oxidized form of HMGB1. Titration of total specific antibodies for the different forms of HMGB1 was made in the serum of 29 HD and 68 HIV-infected patients. Groups of patients were defined according to their clinical neurological status (group 1: stage 1, no HAND, n=22; group 2: stage 2, no HAND with single deficit, n=26; group 3: stage 3, ANI, n=12; group 4: stage 4, MND+HAD, n=8). Statistical significance (two-sided Mann-Whitney test) is shown:*p<0.05, **p<0.01.

(18) FIG. 17. Comparison of levels of IgG specific antibodies for the different forms of HMGB1 in healthy donors (HD) and HIV.sup.+ patients with neurological disorders. Concentrations of IgG antibodies (ng/ml) specific for the different redox forms of HMGB1 (all-thiol, disulfide, oxidized) and specific for Box B were determined in sera from 10 healthy donors (HD) and 103 HIV-1-infected patients according to the same protocol as the one described for FIG. 16. Groups of patients were defined according to their clinical neurological status confirmed by Magnetic Resonance Imaging (MRI). Stage 1: no HAND (n=37), stage 2: no HAND with single deficit (n=37); stage 3: ANI (n=16); stage 4: MND+HAD (n=13). Two-sided Mann-Whitney p values are: **p<0.01; ***p<0.001.

(19) FIG. 18. Correlation between serum levels of IgG antibodies specific for the different forms of HMGB1 and immune activation (percentage of CD8+ CD38+ T cells). Spearman correlations between serum levels of antibodies specific for the different redox forms of HMGB1 and specific for Box B and percentages of CD8+ CD38+ T cells were calculated on sera from 73 HIV-infected patients. The coefficient of correlation (r.sup.2) and p value are reported on each graph.

(20) FIG. 19. Correlation between serum levels of antibodies specific for the different forms of HMGB1 and plasma viral load. Spearman correlations between serum levels of IgG antibodies specific for the different redox forms of HMGB1 and Box B and plasma HIV-1 RNA viral load (VL) were calculated on sera of 73 HIV-infected patients. The coefficient of correlation (r.sup.2) and p value are reported on each graph.

EXAMPLES

(21) I. ELISA Assay for the Detection and Quantification of Human Specific Antibodies for Oxidized HMGB1, Disulfide HMGB1, Fully Reduced HMGB1, Non-Oxidizable Chemokine HMGB1 and HMGB1 Box B

(22) The assay was developed in two steps:

(23) 1. Initially, the inventors optimized the experimental conditions coating the ELISA plates with the non-oxidizable HMGB1 form.

(24) 2. Thereafter, the inventors tested the ability of the assay to detect the presence in human serum of anti-HMGB1 specific antibodies for different redox forms of HMGB1 protein.

(25) The following reagents were used: Disulfide HMGB1 (HMGBiotech, HM-122). Structurally, Disulfide HMGB1 has a disulfide bridge between cysteine residues 23 and 45 and a reduced cysteine residue 106. The Disulfide HMGB1 is the natural protein. Fully reduced HMGB1 (HMGBiotech, HM-116) produced in E. coli. This formulation is fully reduced, it is the natural protein. Non-oxidizable chemokine-HMGB1 (HMGBiotech, HM-132), produced in E. coli, it is a mutant protein where all cysteines are replaced with serines, it is resistant to inactivation by ROS.

(26) Recombinant Box B from HMGB1 (HMGBiotech, HM-052), corresponding to the amino acids 89 to 163 of HMGB1, produced in E. coli from an expression plasmid coding for the mammalian sequence, which is totally identical in human and mouse. Anti-HMGB1 monoclonal antibody DPH1.1 (HMGBiotech, HM-901). Mouse anti-HMGB1 antibody 1E6-E10 (Serotec, Ref MCA4045Z). Human IgG from serum (Sigma; reference I2511) are used as standards during ELISA for IgG antibodies detection. Anti-human IgG (Fc specific)-alkaline phosphatase antibody produced in goat (Sigma; Ref A9544). SIGMAFAST™ p-Nitrophenyl phosphate (pNPP substrate) Tablets (Sigma; reference N2770). MicroWell flat-bottom 96-well plates Nunc (VWR international; reference 62409-112).

(27) Antibodies which bind to HMGB1 are known and can be produced by methods well-known in the art. Examples of commercially available anti-HMGB1 antibodies are anti-HMGB1 monoclonal antibody DPH1.1 (HMGBiotech, HM-901) or 1E6-E10 antibody (Serotec, Ref MCA4045Z). These methods include those which produce polyclonal antibodies to HMGB1 and monoclonal antibodies to HMGB1 or to specific fragments of HMGB1. These antibodies are preferably derived from the same species as the subject to which they are administered and recognized or are induced to the HMGB1 of the same species to which they will be administered. These antibodies may have different isotypes, such as IgA, IgG or IgM isotypes. Antibody fragments which bind HMGB1 may also be employed, including Fab, Fab.sub.2, and single chain antibodies or their fragments.

(28) The ELISA assay to quantitate total specific antibodies for different redox forms of HMGB1 was carried out as follows:

(29) Coating of 96-well plates was performed overnight at 4° C. with 3 μg/ml of the different redox forms of HMGB1 or Box B diluted in DPBS (Dulbecco's Phosphate-Buffered Saline). Simultaneously, coating of serial dilutions of human IgG in DPBS was performed to serve as standards. The different forms of HMGB1 used were: all-thiol, disulfide, oxidized (Ox) and the non-oxidizable (Non-Ox) mutant. In order to maintain (or obtain) the reduced state of the protein, DTT was added to the appropriate wells at a concentration of 1 mM. The oxidized form was obtained adding H.sub.2O.sub.2 (100 mM) to the disulfide HMGB1.

(30) Plates were washed four times with DPBS/0.05% (v/v) TWEEN® 20 (washing buffer), using a microplate washer (Atlantis; Oasys). Similar washings were performed after each step of the ELISA assay. Unbound sites were blocked at 37° C. for 2 hours with DPBS/2% (w/v) BSA (saturation buffer).

(31) Serum samples were treated with one volume of 1.5M Glycine (pH 1.85) for 30 min at 25° C. in a water bath, and further kept on ice and diluted with 1.5M Tris, v/v, pH 9.0. 100 μl aliquots of serum samples were then immediately diluted (from 1/3 to 1/6000) in DPBS/0.05% (v/v) TWEEN® 20/1% (w/v) BSA (antibody dilution buffer), distributed on coated plates and incubated for 1 hour 30 min at 37° C.

(32) Goat anti-human IgG alkaline phosphatase-conjugated antibodies were diluted 1/500 in DPBS/0.05% (v/v) TWEEN® 20/1% (w/v) BSA and added for 1 hour at 37° C.

(33) Detection of antigen-specific antibodies was performed after 30 min of incubation at 37° C. with 100 μl pNPP substrate. The reaction was stopped by addition of 100 μl NaOH 3M and the optical density was read with a Tecan plate reader at 405 nm. Concentration of HMGB1- or BOX B-specific antibodies was calculated according to the standard curve obtained from standard immunoglobulin solution absorbance (FIG. 2).

(34) II. Calibration of the ELISA Assay

(35) To develop this assay, different parameters were assessed using non-oxidizable HMGB1-coated plates:

(36) (i) Assessment of the optimal BSA concentration to saturate wells coated with HMGB1: 2% to 5% BSA concentrations were equally efficient.

(37) (ii) Assessment of the optimal anti-IgG-PAL antibody concentrations (secondary antibody) to reveal bound anti-HMGB1 antibodies: 1/500 dilution was chosen, giving the best results in terms of linearity for batch of antibodies in use.

(38) (iii) Assessment of the optimal HMGB1 concentration in coating buffer: concentrations from 2.5 to 5 μg/ml were the most appropriate.

(39) III. Screening for Interference in the Assay

(40) Once these conditions optimized, the inventors tested the assay for interference and reproducibility.

(41) The standard curves with human IgG always showed a coefficient of determination >0.99 and the samples fell within the standard concentrations (FIG. 3A). As expected, an inverse correlation between sample dilutions and optical density (OD) was observed (FIG. 3B). Nevertheless, once the antibodies concentrations were extrapolated using the standard curve, the inventors noticed an interference effect in the assay (FIG. 3C). The inventors therefore tested different conditions to resolve this issue: adding deoxycholate or TWEEN® 20 to the assay buffers and further diluting the samples.

(42) Influence of Deoxycholic Acid on Interference

(43) Adding deoxycholic acid to the coating buffer did not solve the interference problem of the assay, as shown in FIG. 4.

(44) Effect of TWEEN® 20 on Interference

(45) Adding TWEEN® 20 to the washing buffer and to the antibody dilution buffer did not influence the interference problem of the assay (FIG. 5).

(46) Effect of Serum Dilution on Interference

(47) Serial dilutions of the serum revealed a decrease in interference for dilutions over 1/800. Serum dilutions 1/3000 resulted into no antibody detection (FIG. 6).

(48) IV. Acidic Treatment for the Detection of Total Anti-HMGB1 Antibodies in Human Samples

(49) To determine the assay conditions required for testing human biological samples, a series of human sera have been titrated for the presence of HMGB1-specific antibodies, and assuming that [HMGB1-anti-HMGB1 antibody] complexes were present in biological samples, the influence of pre-treatment with Glycine 1.5M, pH 1.85 to dissociate these immune complexes has been assessed. Serum samples have been either untreated or treated with 1.5M Glycine (v/v, pH 1.85) for 30 min at 25° C. in a water bath, and further kept on ice and diluted with 1.5M Tris, v/v, pH 9.0. Samples were then immediately diluted and distributed on coated plates and tested as described above.

(50) Data in FIG. 7 show that total anti-HMGB1 antibodies were barely detected in human sera, unless they were pretreated with Glycine 1.5M to dissociate the immune complexes.

(51) V. Reproducibility

(52) The mean coefficient of variation of replicate samples within a plate was 3.8% (FIG. 8A), and the coefficient of variation of the results obtained testing the same serum in 6 different experiments was 18.5% (FIGS. 8A and 8B).

(53) VI. Effect of DTT on Monoclonal Antibodies Binding:

(54) Depending on its redox state, HMGB1 exhibits different functions. All-thiol HMGB1 can be maintained in the presence of DTT, while the disulfide form can be induced by H.sub.2O.sub.2 treatment. Prolonged exposition to H.sub.2O.sub.2 leads to the oxidized form (FIG. 1).

(55) The gel in FIG. 9A showed the impact of DTT (1 mM) on the different forms of HMGB1: as expected, non-oxidizable HMGB1 was insensitive to DTT, which did not induce any change in the MW, while the all-thiol molecule (chemokine) required DTT to be stabilized in the right form. Disulfide HMGB1 (cytokine) showed a lower MW compared to All-thiol molecule, as expected, which was increased in the presence of DTT, leading to the MW of All-thiol HMGB1. These data revealed the instability of the various forms of HMGB1, unless an appropriate DTT treatment is applied.

(56) In order to validate the assay, the inventors tested the reactivity of two commercialized monoclonal antibodies (DPH1.1 and 1E6-E10) against the different forms of HMGB1. The plates were coated with various forms of HMGB1 (as obtained from HMGBiotech) and 1 mM DTT was added to specific wells to maintain the reduced state of the protein. FIG. 9B shows that DTT added to the antibody dilution buffer prevented antibody binding, while DTT added only during the coating and saturation steps did not affect the antibody binding (FIG. 9C).

(57) VII. Ability of the Assay to Determine the Specificity of HMGB1-Specific Monoclonal Antibodies Towards the Various Redox Forms of HMGB1

(58) To assess the specificity of the assay, the inventors used the murine anti-HMGB1 antibody 1E6-E10, and first assessed to which form of HMGB1 it was directed. The inventors showed for the first time that 1E6-E10 binds to the non-oxidizable form (Non-ox form), in the presence or absence of DTT, but its binding to the all-thiol form was only detected in the presence of DTT (FIG. 10). This observation confirmed the instability of the reduced form of HMGB1 (as shown on the gel in FIG. 9A), which needs DTT to be stabilized in a reduced state. 1E6-E10 did not recognize the disulfide form, but it did recognize the disulfide form in the presence of DTT, which corresponded to the reduced form (as shown on the gel on FIG. 9A). It did not recognize the oxidized (ox) form either.

(59) Altogether these data indicated that the assay allowed the determination of the specificity of HMGB1 antibodies towards various redox forms of HMGB1, provided they were kept in the right redox state.

(60) VIII. Effect of DTT on Detection of Anti-HMGB1 Antibodies in Healthy Donors (HD)' Sera

(61) Assuming that the mutant, non-oxidizable (Non-ox) form of HMGB1 was not subject to conformational changes following DTT-treatment, the quantification of specific antibodies for the Non-ox form of HMGB1 in human sera was carried out in the presence or absence of 1 mM DTT in both the coating and saturation buffers, in order to evaluate a possible effect of DTT on antibody binding. Indeed, DTT proved to have a very strong effect on the binding of the natural antibodies present in human serum (FIG. 11).

(62) The same experiment was therefore repeated using coating and saturation buffers containing different concentrations of DTT (0.5 mM, 1 mM and 5 mM) (FIG. 12). The amount of natural anti-HMGB1 antibodies present in 3 HD sera was estimated in different experimental conditions. The plate was coated with the non-oxidizable HMGB1 mutant in the presence of different concentrations of DTT. In selected samples, DTT was added at the same concentrations in the saturation buffer. HD samples were tested at a 1/1000 dilution. The presence of DTT in the saturation buffer had a strong effect on the antibody binding. Adding DTT at different concentration just to the coating buffer had a smaller impact on the assay results.

(63) Given the results obtained, the following conditions were chosen for subsequent experiments: 0.5 mM DTT in the coating buffer and no DTT in the saturation buffer.

(64) IX. Detection and Quantification of Anti-HMGB1 Specific Antibodies for Different Redox Forms of HMGB1 and for the HMGB1 Segment Box B in Sera from Healthy Donors (HD) and HIV.sup.+ Patients

(65) The assay was used to titrate serum anti-HMGB1 antibodies against the different forms of HMGB1 in sera from healthy donors (FIG. 13). These data showed for the first time that human sera from healthy donors contained specific antibodies for all redox forms of HMGB1: all-thiol, disulfide and oxidized. Human sera from healthy donors also contained specific antibodies for the protein segment Box B. Variable levels of these antibodies were detected in the five sera tested.

(66) Data in FIG. 14 confirm that sera from all healthy donors contained high levels of specific antibodies for the three redox forms tested and for the HMGB1 segment Box B. 29 healthy donors were tested, and mean concentrations of antibodies were the following:

(67) Anti-Box B: mean 2834 ng/ml [mini 1378-maxi 8126]

(68) Anti-all-thiol: mean 3447 ng/ml [mini 1636-maxi 11697]

(69) Anti-disulfide: mean 3238 ng/ml [mini 1652-maxi 15603]

(70) Anti-oxidized: mean 4825 ng/ml [mini 2596-maxi 8831]

(71) The inventors then addressed the question of the levels of antibodies against the various redox forms of HMGB1 in the context of chronic HIV infection. The group of patients analyzed (n=68) was part of a cohort of 105 chronically HIV-infected patients, classified according to neurological disorders. Group 1 included HIV-1-infected patients without neurological disorders (stage 1: no HAND), whereas groups 2 (stage 2: no HAND with single deficit), 3 (stage 3: Asymptomatic Neurological Impairment—ANI) and 4 (stage 4: Mild Neurological Disorders and HIV-Associated Dementia—MND and HAD) included patients with increasing neurocognitive disorders. FIG. 14 shows that HIV-1 infection was associated with a statistically significant increase in the levels of specific antibodies for all-thiol and oxidized HMGB1. HIV-1 infection was also associated with a statistically significant increase in the levels of specific antibodies for Box B. In contrast, no difference was found in the levels of anti-disulfide antibodies when HIV patients were compared to healthy donors.

(72) X. Quantification of Anti-HMGB1 Specific Antibodies for Different Redox Forms of HMGB1 and for the HMGB1 Segment Box B in Patients with HIV-Associated Neurological Disorders (HAND)

(73) FIG. 15 shows the comparative levels of serum antibodies directed against the redox forms of HMGB1 and against the HMGB1 segment Box B in the patients classified in the 4 groups as described above according to their clinical neurological alterations. Strikingly, the levels of specific antibodies for the oxidized form were significantly increased as compared to the levels of specific antibodies for disulfide, all-thiol or Box B, and this was observed for all groups of subjects, including healthy donors. For group 4 (stage 4) a trend was observed but it did not reach statistical difference, probably due to the low number of subjects in this group.

(74) XI. Specific Antibodies to Oxidized HMGB1 are a Biomarker of HIV Infection and HAND

(75) Concentrations of IgG antibodies (ng/ml) specific for the different redox forms of HMGB1 (all-thiol, disulfide, oxidized) and specific for Box B were determined in sera from healthy donors (HD) and HIV-1-infected patients grouped in stage/group 1 (no HAND), stage/group 2 (no HAND with single deficit), stage/group 3 (ANI—asymptomatic neurological impairment), and stage/group 4 (MND+HAD—Mild Neurological Disorders and HIV-Associated Dementia).

(76) In FIGS. 16 and 17, the levels of antibodies specific for the different forms of HMGB1 were compared between the 5 groups of subjects.

(77) In a first stratification of patients made according to their clinical neurological status (FIG. 16), no statistically significant differences were detected between the 5 groups for specific antibodies to all-thiol form of HMGB1 and for specific antibodies to Box B. The only antibodies that discriminated between patients and healthy donors were specific of the disulfide form of HMGB1 (Two-sided Mann-Whitney p value<0.01 for group 2) and of the oxidized form of HMGB1 (Two-sided Mann-Whitney p value<0.01 for groups 1 and 2 and p<0.05 for group 3).

(78) A second stratification of patients was based on clinical neurological alterations confirmed with Magnetic Resonance Imaging (MRI) providing a more precise definition and distribution of the group of patients (FIG. 17). Using this medical imaging technology, some patients were reallocated a different stage of neurological impairment, which affected the obtained results against the first stratification of patients (FIG. 16). In this second stratification of patients (FIG. 17), the levels of IgG antibodies specific for the various forms of HMGB1 were statistically different between HD and stage 1, stage 2 patients (Two-sided Mann-Whitney p values <0.01 for anti-Box B Abs, anti-all-thiol Abs and anti-disulfide Abs, and two-sided Mann-Whitney p value <0.001 for anti-oxidized Abs). Thus, the levels of antibodies specific for the various forms of HMGB1 allowed the identification of very early stages of neurological impairment (stages 1 and 2). Moreover, anti-oxidized HMGB1 antibodies were kept elevated in stage 3 patients (two-sided Mann-Whitney p value<0.001), making antibodies specific for the oxidized form of HMGB1 the most robust biomarker of HAND stage of HIV infection.

(79) Altogether, these observations should help to address the question of the distribution of these antibodies in pathological conditions, and the correlates with clinical evolution.

(80) XII. Specific Antibodies to Disulfide and Oxidized HMGB1 Positively Correlate with Two Other Biomarkers of HIV Infection

(81) Whole blood from a cohort of 73 chronically HIV-infected patients was tested for the expression of the activation markers HLA-DR and CD38. Blood samples were stained within 8 h of blood draw. CD38 and HLA-DR expression was measured on CD4 and CD8 T cells by six-color flow cytometry using a whole blood cell procedure and monoclonal antibodies specific for CD3 coupled to fluorescein isothiocyanate (FITC), CD8 coupled to peridinin-chrorophyll-protein-cyanin 5.5 (PerCP-Cy5.5), CD4 coupled to phycoerythrin cyanin 7 (PC7), CD45 coupled to allophycocyanin 7 (APC-Cy7) and CD38 coupled to phycoerythrin (PE) and HLA-DR coupled to allophycocyanin (APC). Flow cytometric acquisition and analysis were performed on a FACSCanto flow cytometer and analysis was performed using FACSDiva software. Immune lymphocyte activation is shown by the increased expression of CD8.sup.+ T cells expressing the activation marker CD38. Serum from these patients were tested for antibodies specific for the various forms of HMGB1. Spearman correlations (significance with p<0.05) show that the levels of anti-disulfide and anti-oxidized HMGB1 antibodies (ng/ml) were positively correlated with the % CD8.sup.+CD38.sup.+ T cells, a biomarker of generalized immune activation which characterize chronic HIV infection (FIG. 18).

(82) In the same cohort of 73 chronically HIV-infected patients, quantification of plasma HIV-1 RNA viral load (VL) was performed by RT-PCR (Ampliprep/CobasTaqman Roche Molecular system), with a lower detection limit of 40 copies/ml (1.6 log.sub.10/ml). Some of the patients had undetectable VL. Spearman correlations between viral load and the concentrations of IgG antibodies (ng/ml) specific for the various forms are shown. Positive correlations (significance with p<0.05) were found for antibodies specific for disulfide and oxidized HMGB1 (FIG. 19).

(83) Overall these findings indicated that IgG antibodies specific for the disulfide and oxidized HMGB1 forms, whose levels were statistically increased in very early stages of HIV-associated neurological impairment, were associated with persistent immune activation/inflammation of the CNS, due to persistent viral expression in the patient, albeit most of them were treated with antiretroviral therapy.