human antiganglioside autoantibodies validation of elisa

8/2/2019 Human Antiganglioside Autoantibodies Validation of ELISA

1/14

229

Ann. N.Y. Acad. Sci. 1050: 229242 (2005). 2005 New York Academy of Sciences.doi: 10.1196/annals.1313.024

Human Antiganglioside Autoantibodies

Validation of ELISA

MEPUR H. RAVINDRANATH, SAKUNTHALA MUTHUGOUNDER,

THIRUVERKADU S. SARAVANAN, NAFTALI PRESSER,

AND DONALD L. MORTON

Laboratory of Glycoimmunotherapy, John Wayne Cancer Institute,

Santa Monica, California, USA

ABSTRACT: Gangliosides have a hydrophilic sugar chain that contains antigenic

determinants and a hydrophobic ceramide. In humans, gangliosides elicit a T-cell independent IgM response; antiganglioside IgM autoantibodies may bepentameric or polymeric. A correlation between specific neuropathies and anti-

ganglioside autoantibodies has been confirmed. Although many neurologistsattempt to lower titers of antiganglioside autoantibodies, oncologists are devel-oping strategies to augment production of IgM antibodies that will removeimmunosuppressive gangliosides from the circulation of patients and targetgangliosides and kill tumor cells. Antiganglioside IgM antibodies can cause leak-age of the bloodnerve barrier in a concentration-dependent and complement-

independent manner, bind to neuronal gangliosides to create a neuromuscularblock and serve as a marker of axonal damage in neuropathies such as multiplesclerosis. They are also a promising biomarker of early prostate cancer. Thereis a need to validate the protocol for enzyme-linked immunosorbent assay

(ELISA) of antiganglioside IgM autoantibodies. This validation must considerthe purity of gangliosides from different commercial sources, the coating ofgangliosides onto a solid matrix in a manner that maximizes exposure ofoligosaccharide epitopes to IgM paratopes, techniques to minimize back-ground noise and eliminate nonspecific antibody binding, and carefully definedpositive and negative controls. The validated protocol also must include a

simple formula to estimate titers for several replicas. Finally, antibody titersmust be converted to natural logs for statistical appraisal.

KEYWORDS: antiganglioside; autoantibodies; ELISA; gangliosides; human;

titer; validate; tumor; antigen

Gangliosides are amphophilic molecules with atomic mass units (AMU) ranging

from 1300 to 3000. They have a hydrophilic sugar chain with one or more sialic acids

and a hydrophobic ceramide with sphingosine and a long-chain fatty acid.1,2 Diver-

sity in the gangliosides is created by the number and nature of their sugars, the num-

ber and length of glycosidic linkages of sialic acids, the number of double bonds, and

the extent of fatty-acid hydroxylation. Gangliosides are expressed on the outer layer

Address for correspondence: Mepur H. Ravindranath, Laboratory of Glycoimmunotherapy,John Wayne Cancer Institute, 2200 Santa Monica Boulevard, Santa Monica, CA 904042302.Voice: 310-4495263.

[email protected]


2/14

230 ANNALS NEW YORK ACADEMY OF SCIENCES

of the bilayered lipid membrane of every human cell. Neural tissues and malignant

cells overexpress gangliosides. In aqueous media, the molecules aggregate and form

irregular micelles. In ethanol, micelle formation requires a higher critical micellar

concentration (CMC). Therefore, ethanol suspension is employed to coat poly-

styrene plates in an enzyme-linked immunosorbent assay (ELISA) to measure anti-

ganglioside antibodies.3 During drying in vacuo, CMC increases, causing

attachment of the tail group to the polystyrene plates.The antigenic determinants of common gangliosides are the sugar chains on the

lactosylceramide backbone. TABLE 1 illustrates some of the most common antigenic

determinants of gangliosides in neural, extraneural, and malignant human cells. The

epitope is the antigenic determinant that binds to the paratope, that is, the comple-

mentary combining site (CCS) of an antibody. The upper limit of epitope size is de-

termined by the variable region of the CCS; it is 6 sugar residues for antiganglioside

antibodies. The contact areas involve 25.5 nm of sugar residues and 30.4 nm of

paratope. Fifteen amino acids of the antibody establish 90 van der Waals forces and

9 hydrogen bonds.4 Antigenantibody interactions also may involve salt linkages.

Gangliosides elicit consistent and specific humoral responses without T-cellhelp.58 They do not elicit specific or consistent cellular immune responses. Al-

though protein or peptide antigens are processed by intracellular enzymes and major

histocompatibility complex (MHC) molecules, which transport antigen fragments to

the cell surface and present them to T cells as well as to B cells to generate primary

and secondary antibody responses, gangliosides are notpresented in the context of

MHC molecules, and they do not require T-cell help.8 Oligosaccharide residues of

gangliosides are unable to bind into the groove of MHC molecules, which prevents

direct and specific recognition by conventional specific T cells.57 Because ganglio-

side antigens fail to induce a memory response in humans, antiganglioside antibodies

are invariably IgM, although antiganglioside IgG (commonly IgG2a or IgG3) anti-bodies are found in mice and rabbits. In our ELISA system, we rarely encounter IgG

antibodies to gangliosides in the sera of patients who have been immunized with

ganglioside-based vaccines. To our knowledge, there is no evidence for isotype

switching of antiganglioside IgM or any atypical memory responses mediated by T

TABLE 1. Structure of gangliosides found in normal and malignant human tissues

Glycolipid Structure AMU

GM3 NeuAc2,3Gal1,4Glc1Cer 1236

GM2 GalNAc1,4(NeuAc2,3)Gal1,4Glc1Cer 1385

GD3 NeuAc2,8NeuAc2,3Gal1,4Glc1Cer 1545

GD2 GalNAc1,4(NeuAc2,8NeuAc2,3)Gal1,4Glc1Cer 1694

GM1a Gal1,3GalNAc1,4(2,3NeuAc)Gal1,4Glc1Cer 1547

GM1b NeuAcGal1,3GalNAc1,4Gal1,4Glc1Cer 1547

GD1a NeuAcGal1,3GalNAc1,4(2,3NeuAc)Gal1,4Glc1Cer 1838

GD1b Gal1,3GalNAc1,4(NeuAc2,8NeuAc2,3)Gal1,4Glc1Cer 1838

GD1c NeuAc2,8NeuAc2,3Gal1,3GalNAc1,4Gal1,4Glc1Cer 1838

GT1b NeuAc2,3Gal1,3GalNAc1,4(NeuAc2,8NeuAc2,3)Gal1,4Glc1Cer 2144


3/14

231RAVINDRANATH et al.: HUMAN ANTIGANGLIOSIDE AUTOANTIBODIES

cells. Any IgG antiganglioside antibodies reported in humans could be anti-anti-

idiotypic antibodies or antibodies directed against peptides that mimic ganglioside

epitopes.

Antiganglioside IgM antibodies may be pentamers, hexamers, heptamers, or

other polymers. Unlike conventional pentameric IgM, polymeric IgM may not have

a J chain.9 The polymeric IgM is potentially useful in cancer patients who develop

high titers of antiganglioside antibodies, because antibodies without a J-chain fix

complement 20-fold more efficiently than do conventional pentamers.10 Accruing

evidence suggests that antiganglioside IgM may be secreted by a separate class of B

cells that express the CD5 T-cell marker.11 Although earlier studies suggested that

exogenous adjuvants were necessary to induce an antibody response to ganglioside

antigens,12 our recent study of antiganglioside responses to cryosurgical ablation of

liver metastases indicates otherwise: exogenous adjuvants are not required to elicit

an antiganglioside IgM response.13

In this study of patients with advanced coloncancer, we found that cryoablation of metastatic lesions caused tumor necrosis;

necrotic cells released tumor gangliosides into circulation, provoking a specific anti-

ganglioside antibody response. The increase in antibody levels was followed by a de-

crease in serum gangliosides, suggesting that gangliosides released during necrosis

induced antiganglioside IgM without any exogenous adjuvants. It is possible that ne-

crosis would have acted as a natural adjuvant or would release endogenous adjuvants

such as heat shock proteins to stimulate the ganglioside-specific antibody response.

A disease-specific correlation between specific neuropathies and antiganglioside

antibodies has been confirmed.14,15 Although many neurologists attempt to lower

the titer of the antiganglioside antibodies in neuropathies, oncologists are develop-ing strategies to boost antiganglioside IgM responses to tumor-associated ganglio-

sides. Some antiganglioside IgM antibodies can increase the permeability of the

bloodnerve barrier in a concentration-dependent and complement-independent

manner,16 and some may bind to neuronal gangliosides to create a neuromuscular

block.1720 While antiganglioside IgM can serve as a marker of axonal damage in

neuropathies, including multiple sclerosis, it is far from clear whether the antibodies

cause axonal damage or are a result of axonal damage.21 However, the possible

pathogenesis induced by some of the antiganglioside IgM antibodies should caution

oncologists against indiscriminate boosting of the antiganglioside IgM response.

The development of clinically effective ganglioside vaccines against specific cancersrequires biochemical and immunochemical definition of gangliosides associated

with specific tumor types. Researchers also must monitor changes in the profile of

antiganglioside IgM antibodies during different stages of cancer. This is important to

understand and to identify the homeostatic mechanism by which the host eliminates

gangliosides that are recognized by the immune system as danger signals.

We recently identified a significant antiganglioside antibody response to early-

stage (organ-confined) prostate cancer,22 and our unpublished observations revealed

antiganglioside IgM antibodies in serum and ascites of patients with epithelial

ovarian cancer. These findings suggest that glyco-immunomic studies may lead to the

development of specific antiganglioside IgM as early biomarkers of human cancer.Many laboratories carry out antiganglioside IgM assays without analyzing the

rationale and suitability of each step of the assay system. There is a need to validate

the assay protocol for antiganglioside IgM antibodies. Without proper validation of the

assay to monitor the antiganglioside IgM, antiganglioside IgM biomarkers may not be


4/14


identified or introduced into clinical laboratories. No biomarker assay can be proposed

for clinical use unless it is properly and rigorously validated. Our first publication in

this direction appeared a decade ago.3 Since then, we have developed a standard

operating procedure and validated the ELISA for antiganglioside IgM antibodies.

This ELISA for antiganglioside IgM antibodies has a wide range of clinical and

investigative applications in neoplastic and nonneoplastic disease. In cancer pa-

tients, it can be used to monitor endogenous immune responses to early or localized

disease, evaluate danger signals corresponding to very early stages of disease and

ascertain their correlation with prognosis, assess spontaneous and therapeutically

induced regression, and check the level of tumor necrosis in situ. In neuropathic dis-

ease, it may detect autoimmune neuropathies associated with peripheral neuropa-

thies, celiac disease, Guillain-Barr syndrome, amyotrophic lateral sclerosis, Miller-

Fischer/Bickerstaffs encephalitis, or multiple sclerosis. Further, it can be a tool for

monitoring rheumatoid arthritis, diabetes mellitus I and II, epilepsy, atherosclerosis,Chagas disease, abortion, and stroke. In infectious disease, the ELISA can be used

to check sequelae associated with Campylobacter jejuni, Helicobacter pylori,

cytomegalovirus, HIV, amoebiasis, vaccinia, rotavirus, Mycoplasma pneumoniae,

and neuroborreliosis. Finally, the ELISA can be used to determine the specificity of

antiganglioside monoclonal antibodies and to develop microchip arrays of anti-

ganglioside IgM for human diseases.

VALIDATION OF PROTOCOL FOR ANTIGANGLIOSIDE

ANTIBODY ELISA

Since our previous report,3 we repeatedly validated our standard operating pro-

cedures to obtain high-resolution titers of 1 or more antiganglioside IgM antibodies

in sera from patients with cancer or other diseases and in sera from healthy controls.

Our standard operating procedure entails the following:

(1) Characterizing the purity of ganglioside antigens in different batches and

from different commercial sources.

(2) Coating of gangliosides onto microtiter plates.

(3) Blocking to eliminate background noise.

(4) Maximizing epitopeparatope interaction.

(5) Eliminating nonspecific binding of antibodies.

(6) Taking precautions at the final steps.

(7) Establishing positive and negative controls.

(8) Applying a formula to estimate titers for several replicas.

(9) Converting antibody titers to natural logs.

Characterizing the Purity of Ganglioside Antigens in Different Batches

and from Different Commercial Sources

We routinely use high-performance thin-layer chromatography (HPTLC) to ex-

amine the purity of gangliosides stained with resorcinol or with ganglioside-specific

murine monoclonal antibodies. HPTLC is critical because there is no standard

definition of purity for gangliosides from different commercial sources. We screened


5/14


FIGURE 1. Ganglioside purity was assessed by high-performance thin-layer chroma-tography (HPTLC). Chromatograms were stained by resorcinol-HCl and immunostainedwith ganglioside-specific murine monoclonal antibodies (MAbs). (A) Staining of humanand bovine brain GD1b (3 nmol) from 4 different commercial sources (Sigma, Alexis,Advanced Immunochemical, and Calbiochem). Contaminating glycolipids are indicated bythe thin lines on the right. The GD1b from Alexis was free of contamination and hence used

for ELISA. (B) Staining of GD1a (2 nmol) from 4 different commercial sources (Calbiochem,Alexis, Accurate, and Sigma) by resorcinol and antiganglioside monoclonal antibodiesagainst GD1a (GMR17) and GD2 (14.G2a). Resorcinol staining identified a GM1-likecontaminant that was distinct for GD1a from Accurate, less apparent for GD1a from Cal-biochem and Alexis, and not detectable for GD1a from Sigma. Interestingly, the anti-GD2monoclonal antibody 14.G2a identified a contaminant in GD1a from Calbiochem, Alexis,


6/14


murine monoclonal antibodies for their monospecificity using our ELISA protocol,

and we selected a few of these monoclonal antibodies to test the purity of ganglio-

sides on HPTLC.22

FIGURE 1A shows our experience with GD1b purified from human and bovine

brain by Sigma (St. Louis, MO), Alexis USA (San Diego, CA), Advanced Immuno-

chemical (Long Beach, CA), and Calbiochem (San Diego, CA). In the resorcinol-

stained chromatograms, positions of the contaminating glycolipids are indicated by

the thin lines on the right. Our preferred source for GD1b is Alexis because its GD1b

shows no evidence of contamination.

FIGURE 1B shows contamination in purified GD1a obtained from Calbiochem,

Alexis, Accurate Chemical and Scientific (Westbury, NY), and Sigma. Chromato-

grams were stained by resorcinol and by monoclonal antibodies for all gangliosides

including GD1a (GMR17). Resorcinol staining showed a distinct GM1-like contam-

inant in the preparation from Accurate. This contaminant was less apparent in GD1afrom Calbiochem and Alexis and was not detectable in GD1a from Sigma. Staining

with MAb 14.G2a, the monoclonal antibody for GD2, identified a contaminant at the

position of GD2 in the preparations from Calbiochem, Alexis, and Sigma. Because

the contaminant level was very low in Sigma GD1a and because this GD1a did not

contain the GM1-like contaminant, we chose GD1a from Sigma. Our second choice

was Calbiochem.

FIGURE 1C shows that GD3 from Alexis, Calbiochem, and Sigma was free of

contaminants stainable by either resorcinol or anti-GD3 monoclonal antibodies. We

selected GD3 from Calbiochem because Sigma stopped its supply.

Unfortunately, no commercially obtained GD2 is free of contamination of otherganglioside fractions. GD2 from Advanced Immunochemical (which is reasonably

priced) contains alkali-susceptible GD2, which could be GD2-lactone or O-acetylated

GD2. If it is O-acetylated GD2, 1 or both sialic acids in the GD2 can be O-acetylated.

Base treatment with ammonia (3N) may remove both lactone and O-acetylated

derivatives.

FIGURE 1D shows the result of exposing GD2 to ammonium hydroxide. The scale

in the figure shows the position of the O-acetylated GD2 in the untreated prepara-

tion. Base treatment for 2 h or 4 h removed the top fraction, which would correspond

to lactones or O-acetyl groups in the terminal sialic acid of GD2. However, base

treatment also introduced a new fraction midway between GD2 and the top O-acetylGD2 band, which by its position could be GD2 with an O-acetyl group in the inner

sialic acid (FIG. 1D). Additional base treatment with 7 N ammonium hydroxide for

and Sigma. However, because this contaminant level was very low in Sigma GD1a andbecause this GD1a did not contain GM1-like contaminant, we selected GD1a from Sigmafor ELISA. (C) GD3 from Alexis, Calbiochem, and Sigma was free of contaminants bystaining with resorcinol and monoclonal antibodies. We selected GD3 from Calbiochem forELISA; Sigma no longer produces GD3. (D) Two-dimensional chromatogram of GD2 fromAdvanced Immunochemical shows 2 forms ofO-acetyl GD2 before and after treatment with

ammonium hydroxide at different time intervals. Treatment of GD2 with 2.5 N ammoniumhydroxide de-O-acetylated the terminal sialic, but the O-acetyl group of the internal sialicacid was deacetylated only by treatment with 7 N ammonium hydroxide overnight. (E) Twodifferent lots of GD2 from Advanced Immunochemical. The old lot shows base-labile con-taminants, including O-AcGD2; the new lot is totally devoid of the contaminants after strongbase treatment.


7/14


2 h failed to remove the fraction, but the fraction is eliminated entirely after over-

night treatment in 7 N ammonium hydroxide. The new lot provided by Advanced

Immunochemical represents the final product (FIG. 1E, new lot).

The importance of base treatment before assessing anti-GD2 titers is shown in

FIGURE 1E, which compares 2 lots of GD2 from Advanced Immunochemical. In the

first (early) lot, a major fraction of GD2 had both sialic acids O-acetylated. The

second (new) lot is free of the contamination and is used for routine analyses of sera

for anti-GD2 IgM.TABLE 2 shows the reactivity of sera from 40 patients with stage III melanoma

against (1) GD2 (Advanced Immunochemical) treated in the laboratory with 2.5 N

ammonium hydroxide for 4 h (FIG. 1D) and (2) new lot of GD2 (FIG. 1E). Base treat-

ment abolished the upper band but not the lower band, which may contain O-acetyl

group in the inner sialic acid. Base-treated gangliosides show significantly (P< .02)lower values than untreated, suggesting that the sera may contain IgM antibodies to

O-acetyl GD2. The new lot showed significantly (P< .002) higher values because100% of 3 nmol of new lot is GD2. The lower values obtained with the base-treated

old lot are due to the presence of 80% or less of GD2/well.

In summary, it is critical to obtain pure ganglioside and to use a precise concen-tration (3 nmol) of ganglioside per well before measuring the titers of antiganglio-

side IgM antibodies in ELISA.

Coating of Gangliosides onto Microtiter Plates

The way in which the solid matrix of a microtiter plate is coated with ganglioside

will determine how efficiently the gangliosides sugar domains (epitopes) are ex-

posed for immune recognition. In a previous report,3 we coated ganglioside antigens

onto different polystyrene microtiter plates. Optimal plates were those that had a low

reactivity to antibody without gangliosides in the presence of excipient used for

gangliosides. Absorbency higher than 0.100 at a serum dilution of 1/200 is not

suitable.

Gamma irradiation at 40 kRad (Mark 1-30 irradiator, JL Shepherd and Associates)

can lower background reactivity. Although microtiter plates are treated with gamma

TABLE 2. Anti-GD2 IgM values (expressed in absorbency of sera diluted 1/500) in

40 patients with regional (stage III) metastatic melanoma

TreatmentBackgroundcorrection Minimum Maximum Median Mean STD P valuea

None (old lot) No 0.045 1.609 0.228 0.341 0.318

Yes 0.030 1.595 0.195 0.310 0.316

NH4OH (2.5 N

for 4 h)

No 0.038 1.642 0.168 0.294 0.322 .016

Yes 0.031 1.632 0.154 0.278 0.315 .081 (NS)

New lot No 0.035 2.330 0.285 0.489 0.487 .0019

Yes 0.021 2.310 0.261 0.466 0.484 .0011

NOTE: All three batches of GD2 were from Advanced Immunochemical (Long Beach, CA).

aPaired-sample test between untreated and other treatment group or new lot.


8/14


irradiation during manufacture, the dose and duration of treatment are not available.

We selected Falcon 3915 as the best plate for antiganglioside IgM studies; in recent

months, we found that some of the lots of Falcon 3915 have differed markedly in

their background values, possibly because of differential irradiation or lack of qual-

ity control by the manufacturers. Because lot numbers for Falcon 3915 constantly

change, there is a need to test background noise for each lot before analysis of anti-

ganglioside IgM antibodies. In countries where these plates are not available, inves-

tigators should irradiate the plates in a gamma irradiator and ascertain that the

absorbency of the background noise (at dual wavelength, vide infra) is less than 0.1.

The major problems encountered are ganglioside differences in the length of fatty

acids, the number of double bonds, and the degree of hydroxylation. These charac-

teristics depend, in part, on whether the cell is normal or neoplastic; they also depend

on the commercial source (FIG. 1C). We selected 3 nmol as the optimal concentration

per well. We use 200-proof ethanol as an excipient, rather than methanol or chloro-form or buffer, as cited in the literature.3 In contrast to other prevailing methods (U.S.

patent 6,599,756; 20030049692), we attach the fatty acid residues of the ganglioside

to the plate by 24-h vacuum desiccation. Desiccation can be extended up to 1 month

if undisturbed; however, after 1 week, the surface of the wells should be checked

under a dissection microscope for any evidence of peeling. The oligosaccharide resi-

dues should cluster on the plate in a fashion that simulates their appearance on the cell

surface and facilitates maximal attachment of IgM. Empirically, we determined that

3 nmol of ganglioside is the optimal concentration per well and upon coating in vacuo.

Blocking to Eliminate Background Noise

Intermolecular spaces and ganglioside-free zones of the polystyrene plate are

blocked with a buffer that contains 4% human serum albumin (HSA). We strongly

discourage using gelatin, milk protein, bovine serum albumin, or other xenogenic

proteins for blocking or washing, because they introduce various anomalies. Coating

the plates with 4% HSA in phosphate-buffered saline (pH 7.4) for 90 min will block

nonspecific binding of the serum proteins, IgM, and other antibodies.

Clinical-grade HSA is available from Baxter Healthcare Corporation (Glendale,

CA) and Instituto Grifols (Barcelona, Spain). Known concentrations of this HSA

should be subjected to polyacrylamide gel electrophoresis under reducing or nativeconditions to determine the nature and number of contaminating proteins. The

contaminants in some preparations can significantly lower purity. After several such

investigations, we selected 20% HSA manufactured by Instituto Grifols (storedat

room temperature). This HSA is superior in performance and electrophoretic

albumin/nonalbumin ratio to other preparations.

Blocking with HSA yields a baseline or background value (primary negative con-

trol) that should not exceed an absorbency of 0.100 at titer dilution or 4 dilutions

below titer dilution. For example, if the titer is determined to be 6400, the back-

ground noise should be


9/14


particularly IgM cryoglobulins, precipitate during thawing of frozen fluid speci-

mens, measurements based on fluid from the top of the tube will underestimate true

antibody titers. Cryopreserved sera should be extensively vortexed immediately

upon thawing to resuspend IgM cryoglobulins. Most IgM antibodies against major

gangliosides (GM2, GD3, GD1a and GD1b) are not affected by 5 freeze-thaw cycles

(unpublished data).

Sera can be aliquoted for analyses soon after vortexing. The aliquoted sera can be

refrozen or lyophilized using an accelerated freeze-drying procedure. The lyophilized

sera are reconstituted using blocking buffer that contains 4% HSA (Grifols). The

volume of blocking buffer will be the same as the volume of the original serum used for

lyophilizing. Although we do not use lyophilized serum, others have used it to obtain

accurate measurements of antiganglioside titers (personal communication, N. Yuki).

The most critical step required after reconstitution is incubation of aliquoted sera

diluted 1/100 [with 4% HSA (Grifols) in phosphate-buffered saline (PBS) pH 7.4] at37C for 30 min, before further dilution and addition to microtiter wells. Empirically,we determined that this step is critical to recover IgM precipitated in frozen sera.

Serum is further diluted to 1/200, 1/400, 1/800, 1/1600, 1/3200, 1/6400, and 1/12,800

using detergent-free blocking buffer with 4% HSA in PBS. Depending on the nature

of the disease, the serum can be diluted up to 1/204,800.

Diluted sera (100 L) are overlaid onto antigen-coated (or noncoated) microtiterplates either vertically (from rows A to H; i.e., up to 8 dilutions) or horizontally

(from rows 1 to 12; i.e., up to 12 dilutions). For maximal resolution, the plates are

incubated at 37C for 2 h.

Eliminating Nonspecific Binding of Antibodies

We previously published a table to show the nature and concentration of deter-

gents that have been used to eliminate nonspecific binding of serum antibodies.3 We

studied varying concentrations of Tween-20 to optimize the concentration required

to minimize background noise (optical density below 0.100). Wells overlaid with

serum antibody were washed with a buffer containing 0.1% HSA (Grifols) in PBS

(pH 7.4) with 0.1% of Tween-20. If Tween-20 is stored for a long time, it becomes

ineffective for lowering background noise. We do not store Tween-20 longer than

1 month after opening.Wells are washed manually with a multichannel pipette that is also used to

remove the sera. The washing step is repeated five times, either manually or by an

automated washer (BioRad Model 1575 Immunowasher, BioRad, Hercules, CA).

Taking Precautions at the Final Steps

Precautions at the final steps involve conditions to optimize the oxidation of sub-

strates on the solid matrix and the use of a dual wavelength to correct for noise from

the solid matrix. The secondary antibody that is conjugated to serum IgM is a

peroxidase-conjugated rabbit antihuman Fc5IgM suspended in 4% HSA (Grifols)in PBS (pH 7.4). This antibody is diluted 1/5000 in 4% HSA (Grifols) in PBS

(pH 7.4). The incubation time for maximal resolution is 1 h.

Wells overlaid with serum antibody are washed with a freshly prepared buffer

comprising 0.1% HSA (Grifols) in PBS (pH 7.4) containing 0.1% of Tween-20 (not


10/14


stored for more than 1 month). Serum is removed from wells with a multichannel

pipette, and wash buffer is manually added to wells (using 150 L/well). The BioRadwasher removes the buffer and automatically adds and removes the wash buffer five

times. Then, 100 L of substrate [50 mL ofortho-phenylenediamine dihydrochloride(2 25 mg tablets; Sigma) in citrate-phosphate buffer (pH 5.00) with 21 L of 30%hydrogen peroxide (Sigma, used within 1 month after opening the bottle)] is added

to wells. After exactly 45 min of incubation in the dark, enzymatic oxidation of sub-

strate is arrested by 120 L of 6 N H2SO4. Absorbency is measured after 10 min inan ELISA microtiter plate reader at dual wavelength [A490 A650] to correct foranomalies in the background and titer plates. The absorption maximum after adding

6 N H2SO4 is A492; without 6 N H2SO4, it is A405. It is not correct to measure

absorbency at A490 without adding 6 N H2SO4 or at A400 after adding 6 N H2SO4.

Establishing Positive and Negative ControlsPositive or reference controls for antiganglioside IgM vary with the type of dis-

ease or cancer. Serum aliquots are obtained on different days from 1 or more patients

who have high antibody titers; cryopreserved specimens are thawed, pooled, and

vortexed to obtain 10 to 15 mL of sera, which is further split into 10-L aliquots in500-L Eppendorf (color) vials.

Negative controls are obtained by pooling sera from adult males older than

50 years; age is important because the titers of most antiganglioside antibodies are

low in younger patients.14 Again, about 10 to 15 mL of pooled sera is divided into

10-L aliquots in 500-L Eppendorf (colorless) vials.

FIGURE 2. Consistency of antiganglioside ELISA. During validation studies, ELISAis repeated on different days with the same positive and negative control sera. The figureillustrates anti-GM2 IgM values in positive control sera pooled from different specimens ofa patient with AJCC stage III melanoma (top) and negative control sera pooled from healthyvolunteers older than 50 years (bottom). When control values peaked, ELISA failed. For a validassay, the coefficient of variation for positive and negative controls must be less than 15%.


11/14


One vial each of positive and negative control is analyzed simultaneously with

each study specimen. FIGURE 2 shows the values obtained by a single investigator

(SM) who repeated the ELISA daily for 50 days.

Applying a Formula to Estimate Titers for Several Replicas

With Microsoft Excel software, the absorbency values of the microtiter plates can

be calculated by the following procedure (developed by D. Soh with M. H. Ravin-

dranath, unpublished):

After reading a plate, use the copy window (Ctrl + C) command in Softmaxto copy the data in Window. After opening Excel program, paste the data

(Ctrl + V). This procedure is repeated for every plate.

Correct experimental values against background noise by using the subtrac-

tion command [e.g., = (A1 A2)]. To find the titer at 0.1 absorbency(A490nm A650nm), only 2 readings are assessed. For example, in the chartbelow, cell column 1 has a dilution (X) of 1600 and absorbency (Y) of 0.171;

cell column 2 has a dilution of 3200 and an absorbency of 0.082. Cell column

B and row 3 in Excel (slope value) is used (clicked) to determine slope

(change in Y, change inX). After clicking B3, type in the cell = slope (, andhighlight cells B1 and B2 (Yvalues), insert a comma, highlight cells A1 and

A2 (Xvalues), and close parenthesis. The slope should always be negative,

and there should always be a comma between theXand Yvalues:

The titer at 0.1 is calculated by the following formula: (0.1 lowest absor-

bency value)/slope) + highest dilution factor(which corresponds to the lowestabsorbency value). Click on B4 cell; type = ((0.1 ; highlight B2 (lowestabsorbency value) and close parenthesis; type / and highlight B3 (slope);

Type ) + and then highlight A2 (highest dilution factor); likewise in thechart: = ((0.1 B2)/B3) + A2. Round up the titer value by changing the num-ber of decimals to 0. Highlight the B4 cell and right click. In the menu, choose

format cells and click on number; then change the decimal value to 0:

Cell column X Y

1 1600 0.171

2 3200 0.082

3 Slope 0.000055625

4 Titer

Cell column A B

1 1600 0.1710

2 3200 0.082

3 Slope 0.000055625

4 Titer 2876.404494


12/14


If the copy-and-paste function is used to find slope and titer values, the 2 dilu-

tion values must be above and below 0.1. The same procedure also can be

used to find titers at a different absorbency such as 0.2, but the titer equation

must be changed to 0.2 and all other dilution and absorbency values must be

above and below 0.2.

Converting Antibody Titers to Natural Logs

Comparison of the antiganglioside IgM antibody titers for different gangliosides

and in relation to treatment or stages of disease cannot use mean or median values

because of the large standard deviation. To overcome this problem, we convert

values to natural log titers. TABLES 3A and 3B summarize the log titers of different

gangliosides obtained from healthy volunteers and patients with benign prostatehyperplasia, organ-confined prostate cancer (stage T1/T2), and unconfined prostate

cancer (stage T3/T4). Further elaboration of the same data with an expanded sample

size is shown elsewhere.23

TABLE 3A. ANOVA assessment ofP values showing significant differences in log

titers of anti-GD3, anti-GD2, and anti-GD1a IgM antibodies in sera from healthycontrols and patients with benign prostatic hyperplasia (BPH) or prostate cancer

(CaP)

IgMtarget

Healthy(n= 11)

BPH(n= 10)

T1/2 CaP(n= 20)

T3/4 CaP(n= 7)

ANOVAP

valueMean

SD MedianMean

SD MedianMean

SD MedianMean

SD Median

GM1 4.48 1.26

4.61 4.61 0.00

4.61 5.28 0.99

4.61 5.46 1.17

5.70

GM2 5.93 0.65

5.97 5.16 0.99

4.61 5.16 0.98

4.61 6.07 1.55

5.99 .073

GM3 4.43 0.22

4.61 4.68 0.22

4.61 5.19 1.19

4.61 5.15 1.25

4.61 .181

GD3 5.45 0.91

5.19 4.61 0.00

4.61 4.78 0.86

4.61 4.51 0.27

4.61 .022

GD2 4.18 1.01

4.61 5.06 0.77

4.61 5.62 1.19

5.70 5.74 1.37

6.21


13/14


REFERENCES

1. ANDO, S. 1983. Gangliosides in the nervous system. Neurochem. Int. 5: 507537.2. WIGAND, H. 1985. Gangliosides. New Compr. Biochem. 10: 199260.3. RAVINDRANATH, M.H., R.M. RAVINDRANATH, D.L. MORTON & M.C. GRAVES. 1994.

Factors affecting the fine specificity and sensitivity of serum antiganglioside anti-bodies in ELISA. J. Immunol. Methods 169: 257272.

4. RICH, R., R. FLEISHER, T. A. SCHWARTZ, et al. 1996.In Clinical Immunology: Principles

and Practice. Mosby. St. Louis.5. HARDING, C.V., J. KIHLBERG, M. ELOFSSON, et al. 1993. Glycopeptides bind MHC

molecules and elicit specific T cell responses. J. Immunol. 151: 24192425.6. HARDING, C.V., R.W. ROOF, P.M. ALLEN & E.R. UNANUE. 1991. Effects of pH and

polysaccharides on peptide binding to class II major histocompatibility complexmolecules. Proc. Natl. Acad. Sci. USA 88: 27402744.

7. ISHIOKA, G.Y., A.G. LAMONT, D. THOMSON, et al. MHC interaction and T cell recognitionof carbohydrates and glycopeptides. J. Immunol. 148: 24462451.

8. FREIMER, M.L., K. MCINTOSH, R.A. ADAMS, et al. 1993. Gangliosides elicit a T-cellindependent antibody response. J. Autoimmun. 6: 281289.

9. RANDALL, T.D., L.B. KING & R.B. CORLEY. 1990. The biological effects of IgMhexamer formation. Eur. J. Immunol. 20: 19711979.

10. DAVIS, A.C., K.H. ROUX & M.J. SHULMAN. 1988. On the structure of polymeric IgM.Eur. J. Immunol. 18: 10011008.

11. RAVINDRANATH, R.M., M.H. RAVINDRANATH & M.C. GRAVES. 1997. Augmentation ofnatural antiganglioside IgM antibodies in lower motor neuron disease (LMND) androle of CD5+ B cells. Cell Mol. Life Sci. 53: 750758.

12. LIVINGSTON, P.O. 1995. Approaches to augmenting the immunogenicity of melanomagangliosides: from whole melanoma cells to ganglioside-KLH conjugate vaccines.Immunol. Rev. 145: 147166.

13. RAVINDRANATH, M.H., T.F. WOOD, D. SOH, et al. 2002. Cryosurgical ablation of livertumors in colon cancer patients increases the serum total ganglioside level and thenselectively augments antiganglioside IgM. Cryobiology 45: 1021.

14. RAVINDRANATH, M.H., A.M. GONZALES, K. NISHIMOTO, et al. 2000. Immunology of

gangliosides. Indian J. Exp. Biol. 38: 301312.15. GALLARDO, E., R. ROJAS-GARCIA, R. BELVIS, et al. 2001. Antiganglioside antibodies:

when, which and for what? Neurologia 16: 293297.16. KANDA, T., T. IWASAKI, M. YAMAWAKI, et al. 2000. Anti-GM1 antibody facilitates leakage

in an in vitro blood-nerve barrier model. Neurology 55: 585587.17. WINER, J.B. 2001. Guillain Barre syndrome. Mol. Pathol. 54: 381385.

TABLE 3B. Pairwise comparison by least significant differences method revealing

that anti-GD1a IgM log titers distinguish between T1/2 and T3/4 CaP

Pairwise comparisons

P values

Anti-GD3 Anti-GD2 Anti-GD1a

Healthy vs. BPH .010a .076 .992

Healthy vs. T1/2 CaP .016a .001 .018

Healthy vs. T3/4 CaP .009a .006 .756

BPH vs. T1/2 CaP .546 .205 .021

BPH vs. T3/4 CaP .757 .224 .767

T1/2 CaP vs. T3/4 CaP .398 .804 .019

aLower than healthy volunteers.


14/14


18. KUWABARA, S., K. OGAWARA, K. MIZOBUCHI, et al. 2000. Isolated absence of F wavesand proximal axonal dysfunction in Guillain-Barre syndrome with antigangliosideantibodies. J. Neurol. Neurosurg. Psychiatry 68: 191195.

19. OBI, T., T. MURAKAMI, M. TAKATSU, et al. 1999. Clinicopathological study of an

autopsy case with sensory-dominant polyradiculoneuropathy with antigangliosideantibodies. Muscle Nerve 22: 14261431.20. ARASAKI, K., S. KUSUNOKI, N. KUDO & M. TAMAKI. 1998. The pattern of antiganglio-

side antibody reactivities producing myelinated nerve conduction block in vitro. J.Neurol. Sci. 161: 163168.

21. SADATIPOUR, B.T., J.M. GREER & M.P. PENDER. Increased circulating antigangliosideantibodies in primary and secondary progressive multiple sclerosis. Ann. Neurol. 44:980983.

22. RAVINDRANATH, M.H., S. MUTHUGOUNDER, N. PRESSER, et al. 2004. Gangliosides oforgan-confined versus metastatic androgen-receptor-negative prostate cancer. Biochem.Biophys. Res. Commun. 324: 154165.

23. RAVINDRANATH, M.H., S. MUTHUGOUNDER, N. PRESSER, et al. 2005. Endogenous

immune response to gangliosides in patients with confined prostate cancer. Int. J.Cancer 116: in press.

human antiganglioside autoantibodies validation of elisa

Documents