Protocol for measuring anti-fentanyl antibodies in mouse serum by enzyme-linked immunosorbent assay

Summary

Enzyme-linked immunosorbent assay (ELISA) offers an effective, inexpensive, and reliable approach for the analysis of humoral immune responses. Here, we describe a protocol for measuring anti-fentanyl antibodies generated by the immunization of mice with novel opioid vaccine candidates. We describe steps for coating BSA-fentanyl antigen and standard wells. We then detail procedures for preparing buffers and performing assays. This protocol is adaptable for use with either BALB/c or C57BL/6J mice strains and can measure antibody isotypes to elucidate class switching.

For complete details on the use and execution of this protocol, please refer to Stone et al.¹

Graphical abstract

Highlights

• •This protocol can determine antibody class switching using the indirect ELISA method
• •This protocol detects mouse anti-fentanyl antibodies using a fentanyl-protein conjugate
• •This protocol can test up to 40 mouse samples in duplicate per plate and isotype
• •The protocol has been optimized to allow for the use of either BALB/c or C57BL/6 mice

Publisher’s note: Undertaking any experimental protocol requires adherence to local institutional guidelines for laboratory safety and ethics.

Enzyme-linked immunosorbent assay (ELISA) offers an effective, inexpensive, and reliable approach for the analysis of humoral immune responses. Here, we describe a protocol for measuring anti-fentanyl antibodies generated by the immunization of mice with novel opioid vaccine candidates. We describe steps for coating BSA-fentanyl antigen and standard wells. We then detail procedures for preparing buffers and performing assays. This protocol is adaptable for use with either BALB/c or C57BL/6 mice strains and can measure antibody isotypes to elucidate class switching.

Before you begin

The protocol below describes the steps to measure anti-fentanyl antibodies in serum samples collected from either BALB/c or C57BL/6J mice immunized with vaccines designed to help treat opioid use disorder (OUD) and/or prevent overdose death. BALB/c and C57BL/6J mice strains differ in the IgG2 isotype. BALB/c mice express IgG2a whereas C57BL/6J mice express IgG2c. A standard mouse IgG2a and IgG2c are both listed in the key resources table within this protocol for use with the corresponding mouse strain. Mouse blood is first collected into microcentrifuge tubes and centrifuged at 1500 × g for 7 min. After centrifugation, supernatant is pipetted into a new, clean microcentrifuge tube and centrifuged again at 1500 × g for 7 min. After the second centrifugation, the serum supernatant is transferred to strip tubes and stored at −80°C.

This protocol utilizes the indirect ELISA method for sample detection. The standard curves within this protocol will follow different ELISA methodology depending on the isotype. Pay close attention to the antibody isotype being tested and its corresponding standards when running the assay. Total IgG and IgG1 use the indirect ELISA method with a mouse anti-fentanyl monoclonal antibody. IgG2c and IgA use the direct ELISA method with normal mouse IgG2c or IgA bound directly to the plate at the time of coating. IgG2a uses the capture ELISA method with an anti-mouse IgG2a coated to the plates and the normal mouse IgG2a added at the time of sample addition.

Institutional permissions

All experiments involving animals must be conducted in accordance with all federal, state, and institutional regulations, policies, and procedures including the Institutional Animal Care and Use Committee (IACUC). All experiments conducted in the generation of this protocol were approved by the Boston Children’s Hospital IACUC.

CRITICAL: In many jurisdictions, such as the US, fentanyl (FEN), other opioids, and their derivatives are listed as controlled substances (e.g., under Schedule 2 by the US Drug Enforcement Agency (DEA)²). Accordingly, these substances require special permissions and procedures for handling. Before beginning any work with opioids check with your institutional safety department to ensure all required permits and safety measures are in place prior to beginning experiments.

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Normal mouse IgA – serially diluted 500 ng–0.68 ng	Santa Cruz Biotechnology	Catalog: sc-516590
Goat anti-mouse IgG human adsorbed (ads) (unlabeled) – diluted to 2 μg/mL	Southern Biotech	Catalog: 1030-01; RRID: AB_2794290
Mouse IgG2c (unlabeled) – serially diluted 500 ng–7.81 ng	Southern Biotech	Catalog: 0122-01; RRID: AB_2794064
Normal mouse IgG2a – serially diluted 500 ng–7.81 ng	Santa Cruz Biotechnology	Catalog: sc-3878
Mouse anti-fentanyl IgG – serially diluted 500 ng–7.81 ng	CalBioreagents	Catalog: M571
Rabbit anti-mouse IgG-AKP – diluted 1:500	Sigma-Aldrich	Catalog: A1902; MDL: MFCD00162640
Rat anti-mouse IgG1-AKP – diluted 1:1,000	BD Biosciences	Catalog: 557272; RRID: AB_396612
Goat anti-mouse IgG2a-HRP – diluted 1:5,000	Southern Biotech	Catalog: 1081-05; RRID: AB_2736843
Goat anti-mouse IgG2c-HRP – diluted 1:5,000	Southern Biotech	Catalog: 1078-05; RRID: AB_2794462
Goat anti-mouse IgA-HRP – diluted 1:1,000	Sigma-Aldrich	Catalog: A4789; MDL: MFCD00162618
Chemicals, peptides, and recombinant proteins
Bovine Serum Albumin (BSA)-fentanyl conjugate	CalBioreagents	Catalog: C149
Phosphate-buffered saline (PBS) (10X) DNase, RNase protease free	Fisher Scientific	Catalog: BP399-20
Alkaline phosphatase (AKP) substrate: p-nitrophenyl phosphate, disodium salt (pNPP)	Sigma	Catalog: N2765; MDL: MFCD00066288
Horseradish peroxidase (HRP) substrate:3,3′,5,5′-Tetramethylbenzidine (TMB) substrate reagent set	BD Biosciences	Catalog: 555214; RRID: AB_2869044
Diethanolamine (DEA) buffer	Sigma	Catalog: D8885; MDL: MFCD0002843
Magnesium chloride hexahydrate (MgCl 6H2O)	Sigma	Catalog: M2773; MDL: MFCD00149785
Other
Automatic plate washer	Agilent	Model: BioTek 405TS
Optical plate reader (optional)	Molecular Devices	Model: SpectraMax iD3
Pipetman: 2, 200, and 1,000 μL	Gilson	Catalog: F144054M, F144058M, F144059M
Multichannel pipette: 8 or 12 channel, 200–300 μL	Gilson	Catalog: F144074, F144075
Globe serological pipettes: 5, 10, and 25 mL	Midwest Scientific	Catalog: GLB-SERO-5ML, GLB-SERO-10 mL, GLB-SERO-25ML
pH meter	Fisher Scientific	Model: Accumet AE150
Digital scale	Mettler Toledo	Model: ME103E
96-well Clear flat bottom polystyrene high bind microplate	Corning	Catalog: 9018
150 mL Polystyrene bottles	Thermo Scientific	Catalog: 455-0150
Reagent Reservoirs	VWR	Catalog: 89094-674

Materials and equipment

Wash buffer

Reagent	Final concentration	Amount
10X PBS	1X	200 mL
Deionized Water (diH2O)	N/A	1,800 mL
Tween 20	0.05%	1 mL

Blocking buffer

Reagent	Final concentration	Amount (per plate)
10X PBS	1X	1 mL
diH2O	N/A	9 mL
Tween 20	0.05%	5 μL
Non-fat Dry Milk (NFDM)	2%	0.2 g

Assay buffer

Reagent	Final concentration	Amount (per plate)
10X PBS	1X	1 mL
diH2O	N/A	9 mL
Tween 20	0.05%	5 μL
NFDM	0.2%	0.02 g

DEA buffer

Reagent	Final concentration	Amount
diH2O	N/A	800 mL
Diethanolamine	9.8%	98 mL
Magnesium Chloride Hexahydrate (MgCl 6H2O)	0.1 mg/mL	100 mg
Combine above components then adjust pH to 9.8 and bring final volume to 1 L with diH2O.

Step-by-step method details

Day 1—Coating microtiter plates

Timing: 1 h active

• 1.
  Coating BSA-Fentanyl Antigen.

  • a.
    Determine the number of high binding immunoassay plates needed to test all samples.

    • i.
      Each plate can test a maximum of 40 samples in duplicate against a single antibody isotype. To determine the total number of plates needed, first divide the number of samples by 40, round up to the nearest whole number.
    • ii.
      Second, multiply the number from above (1a) by the number of antibody isotypes to test. This number is the total number of plates needed for the assay.

      Note: It is recommended to make plates as needed as storing plates can lead to increases in plate variance.

  • b.
    Dilute the BSA-Fentanyl Conjugate from CalBioreagents to 2 μg/mL in 1X PBS (prepared from 10X stock using diH₂O).

    • i.
      A volume of 5 mL of 2 μg/mL diluted BSA-Fentanyl Conjugate is needed per plate.
  • c.
    Diluted BSA-Fentanyl conjugate is added at 50 μL/well to the plates as follows:

    Antibody isotype target	Wells containing BSA-FEN	Volume per well
    Total IgG or IgG1	All wells	50 μL
    IgG2a, IgG2c, or IgA	Wells in plate columns 3 through 12	50 μL

    CRITICAL: Ensure the surface of all the wells containing antigen are covered by the coating antigen solution.

    Note: This protocol utilizes fentanyl conjugated to BSA. Other carrier proteins could be used such as tetanus toxoid (TT) so long as the carrier protein is not included in the immunization formulation. The high binding ELISA plates from Costar are capable of binding biomolecules greater than 10,000 kDa. The molecular weight of fentanyl is 0.3365 kDa making it too small to bind to the plates on its own. Conjugation to a biomolecule like BSA or TT is necessary to coat the fentanyl to the ELISA plate.

• 2.
  Coating standard wells for IgG2a, IgG2c, or IgA:

  • a.
    To generate standard curves for mouse IgG2a, IgG2c, or IgA, an unconjugated, normal mouse isotype of each will be used. See Figure 1 for coating strategy.

    • i.
      For IgG2a, dilute stock Goat Anti-mouse IgG Human ads-Unlabeled from Southern Biotech to 2 μg/mL in 1X PBS. Add 50 μL to each well of columns 1 and 2 of the plates for IgG2a.

      Note: The goat anti-mouse IgG Human ads-unlabeled from Southern Biotech has been purified with minimal reactivity to human proteins and cross reacts with mouse IgG1, IgG2a, IgG2b, IgG2c, and IgG3.

    • ii.
      For IgG2c, serially dilute stock Mouse IgG2c Unlabeled from Southern Biotech 1:2 from 500 ng to 7.81 ng in 1X PBS. Add each dilution in duplicate to the IgG2c plates in descending order, 50 μL/well. To the last row add 1X PBS without antibody to serve as a plate blank.
    • iii.
      For IgA, serially dilute stock Normal Mouse IgA from Santa Cruz Biotechnology 1:3 from 500 ng to 0.68 ng in 1X PBS. Add each dilution in duplicate to the IgA plates in descending order, 50 μL/well. To the last row add 1X PBS without antibody to serve as a plate blank.
  • b.
    Plates are covered and incubated at 4°C overnight to allow antigen and standards to bind to the plate.

    Optional: Alternatively, coating can be accomplished by incubating at 25°C for one hour (e.g., in an incubator).

Figure 1.

IgG2a, IgG2c, and IgA coating template

Wells shaded green are reserved for standards (Std) and plate blanks. Wells shaded blue correspond to 2 μg/mL BSA-Fentanyl Conjugate.

Day 2—Running the assay

Timing: 4.5 h

• 3.
  Buffer Preparation.

  • a.
    Prepare all buffers needed for running the assay. Buffer recipes can be found in the materials and equipment section of this protocol.

    • i.
      For blocking buffer and assay buffer, a minimum volume of 10 mL per plate of each is required for the entire assay. More volume may be required for making dilutions.
    • ii.
      For DEA buffer, if not previously prepared, a minimum of 20 mL is required. DEA buffer has a shelf life of one month from the date it was made when stored at 4°C.

      Note: 20 mL is the volume of buffer required to dissolve 1 tablet of pNPP substrate. Dissolve the tablet on the day of use and keep in the dark at room temperature until dissolved and ready to use.

      Note: 20 mL is enough volume to develop 4 plates with alkaline phosphatase conjugated secondary antibody. Increase volume as necessary in increments of 20 mL.

      Note: The pH of DEA buffer may change over time and may cause background values to increase.

    • iii.
      The quantity of wash buffer will depend on the number of plates and type of plate washer used.
• 4.
  Assay Procedure.

  • a.
    Using an automatic plate washer, wash plates 3X with 250 μL/well of wash buffer to remove unbound coating.

    • i.
      If not using an automatic plate washer, fill each well with wash buffer using a squirt bottle. Shake gently and dump into a liquid waste receptacle. Repeat three times and tap gently to remove excess wash buffer.
  • b.
    Add at least 100 μL/well of blocking buffer to each plate.
  • c.
    Incubate plates with blocking buffer for no less than 1 h at room temperature.
  • d.
    Dilute samples and standards:

    • i.
      Dilute mouse serum samples 1:100 in assay buffer. A total of 100 μL of diluted sample per antibody isotype is required. Mix well.
    • ii.
      Dilute the mouse anti-fentanyl monoclonal from CalBioreagents to 500 ng in assay buffer. Then serially dilute the 500 ng dilution 1:2 to create the following dilutions: 500 ng, 250 ng, 125 ng, 62.5 ng, 31.25 ng, 15.6 ng, and 7.81 ng. Final volume of each dilution needs to be equal to no less than 100 μL per plate.

      Optional: If running IgG2a, dilute the Normal Mouse IgG2a from Santa Cruz Biotechnology to 500 ng in assay buffer. Then serially dilute the 500 ng dilution 1:2 to create the following dilutions: 500 ng, 250 ng, 125 ng, 62.5 ng, 31.25 ng, 15.6 ng, and 7.81 ng. Final volume of each dilution needs to be equal to no less than 100 μL per plate.

  • e.
    Using an automatic plate washer, wash plates 3X with 250 μL/well of wash buffer to remove blocking buffer.
  • f.
    Add standards and samples to plates, 50 μL/well in duplicate. Refer to Figure 1 for plate layout.

    • i.
      To the IgG and IgG1 plates, add the mouse anti-fentanyl monoclonal serial dilutions (step 10b) to columns 1 and 2 of the plate(s) descending in concentration by row with row A the highest concentration. To row H, add 50 μL of assay buffer to serve as a plate blank.
    • ii.
      For the IgG2c and IgA standard curve wells, no antibody is added at this step. Instead add 50 μL of assay buffer to prevent the wells from drying out.

      Optional: If running IgG2a, add the diluted Normal Mouse IgG2a (step 11c) to the columns coated with the goat anti-mouse IgG Human ads-unlabeled (step 5a) descending in concentration by row with row A the highest concentration. To row H, add 50 μL of assay buffer to serve as a plate blank.

    • iii.
      Add diluted samples (step 10a) to the empty wells in duplicate, 50 μL per well. If the number of samples requires more than one plate per antibody isotype, denote the plates in a manner to track which samples are on which plate. Repeat for each antibody isotype.
  • g.
    Incubate plates with standards and samples for at least 1 h at room temperature.
  • h.
    Dilute detection antibodies in assay buffer (Table 1). 5 mL per plate required volume.
  • i.
    Using an automatic plate washer, wash plates 3X with 250 μL/well of wash buffer to remove unbound sample and standard.
  • j.
    Add 50 μL of detection antibody to each well of corresponding isotype plate(s).
  • k.
    Incubate detection antibody on plates for at least 1 h at room temperature.
  • l.
    Prepare enzyme substrates.

    • i.
      Remove TMB from 4°C and mix equal parts A and B to create a necessary final volume allowing for 50 μL/well on plates with HRP conjugated detection (IgG2a/IgG2c, IgA).

      CRITICAL: Allow TMB to warm to room temperature before adding to plates.

    • ii.
      Remove DEA buffer from 4°C and add 1 tablet of pNPP per 20 mL of DEA. Allow to dissolve then mix well by inversion or vortex.

      CRITICAL: Allow DEA buffer/pNPP substrate to warm to room temperature before adding to plates and ensure the tablet is completely dissolved.

  • m.
    Using an automatic plate washer, perform a final ELISA wash by washing plates 5X with 300 μL/well of wash buffer to remove unbound detection antibody.

    Note: The extra volume and steps in this wash help ensure no residual enzyme conjugate before addition of substrate.

  • n.
    Add 50 μL of TMB to each well of IgG2a/IgG2c and IgA plates (HRP conjugated detection) and 50 μL of pNPP substrate to each well of IgG and IgG1 plates (AKP conjugated detection).
  • o.
    Incubate substrate on plates for no more than 30 min, in the dark, and at room temperature.
  • p.
    After development, add 50 μL of 2N H₂SO₄ to the wells with TMB to stop the reaction. Add 50 μL of 2N NaOH to the wells with pNPP to stop the reaction.
  • q.
    Read plates immediately on optical plate reader. Plates with TMB and stop solution are read at 450 nm. Plates with pNPP and stop solution are read at 405 nm.
  • r.
    Once plates have been read and data saved, discard the plates following relevant institutional waste handling procedures.
  • s.
    Analyze data collected to determine concentration of antibody isotypes (see Quantification and statistical analysis).

Table 1.

Detection antibody working dilutions

Antibody	Dilution
Total IgG Detection: Rabbit anti-mouse IgG-Alkaline phosphatase (AKP)	1:500
IgG1 Detection: Rat anti-mouse IgG1-AKP	1:1,000
IgG2a Detectiona: Goat anti-mouse IgG2a-Horseradish Peroxidase (HRP)	1:5,000
IgG2c Detectionb: Goat anti-mouse IgG2c-HRP	1:5,000
IgA Detection: Goat anti-mouse IgA-HRP	1:1,000

^aFor use with BALB/c mice.

^bFor use with C57BL/6J mice.

Expected outcomes

Upon reading the plates, wells in the standards should show decreasing optical densities (OD) with the highest optical density correlating with the highest concentration on the standard curve. For this assay, it is possible to achieve a Goodness of Fit (r²) greater than 0.98 for each antibody isotype. Measured values of the unknown samples should fall within the standard curve OD range. Failure to fall within the minimum or maximum values of the standard curve may result in inaccurately calculated, or uncalculated, values. If the values exceed the maximum or minimum value of the standard curve, consider re-running the assay while adjusting the sample dilution to correct for the drift outside the range of the standard curve.

The concentration and dilution of coating antigen and detection antibody were optimized by testing reagents against one another. Standards were created from commercially available reagents of known concentrations, either anti-FEN monoclonal for IgG and IgG1, or unlabeled mouse whole Ig molecules for IgG2a, IgG2c, and IgA. Anti-FEN antibody was tested against concentrations of BSA-FEN (0.5, 1.0, 2.0 μg/mL) coated to a high binding microtiter plate. The mouse anti-FEN monoclonal was diluted as a standard curve (500 ng, 100 ng, 20 ng, 0 ng) and allowed to bind to the antigen. Anti-mouse IgG or IgG1 conjugated to alkaline phosphatase was diluted to various test concentrations, selected based on the manufacturer’s recommended dilutions, and allowed to bind to the mouse anti-FEN antibody. pNPP substrate was reacted for 30 min and then stopped with 2N NaOH. Reactions were read at 405 nm, and standard curves were interpolated using GraphPad Prism version 9.5.1 (Figures 2A and 2B) (See: Quantification and statistical analysis). The combination of coating antigen concentration and detection antibody working dilution which yielded the greatest r² greater than 0.98 was selected for use in the assay.

Figure 2.

Assay optimization curves

Total IgG detection antibody was tested against increasing concentrations of BSA-FEN coatings of (A). The 1:500 dilution of IgG detection antibody with the 2 μg/mL coating antigen yielded the strongest correlation (r² = 0.9953). IgG1 detection antibody was tested against increasing concentrations of BSA-FEN coatings (B). The 1:1000 dilution of IgG1 detection antibody with the 2 μg/mL coating antigen yielded the strongest correlation (r² = 0.9987). IgG2a was coated directly to the ELISA plate at standard curve concentrations (C). No condition for direct IgG2a detection yielded satisfactory correlation coefficients. IgG2a capture antibody was tested at increasing concentrations with various IgG2a detection antibodies (D). When detecting IgG2a standards through the capture method the 2 μg/mL coating antibody with 1:5000 detection gave the strongest correlation coefficient (r² = 0.9983). Direct detection of IgG2c standards with an anti-IgG2c detection antibody was optimized at a detection concentration of 1:5000 (r² = 0.9968) (E). Direct detection of IgA found a detection concentration of 1:1000 yielded the strongest correlation (r² = 0.9989) (F).

IgG2a standards were optimized first by directly coating unconjugated mouse IgG2a directly to the plate at standard concentrations (500 ng, 100 ng, 20 ng, 0 ng). This direct ELISA method however did not yield satisfactory results evidenced by low correlation coefficients (Figure 2C). To correct this, an unlabeled goat anti-mouse IgG cross absorbed against human Ig was coated to the plate. This antibody is capable of cross reacting with mouse IgG1, IgG2a, IgG2b, IgG2c, and IgG3. Cross adsorption to human Ig is a secondary purification technique done to improve the sensitivity of the purified antibody sometimes offered on commercial antibodies. The unlabeled mouse IgG2a was then diluted as a standard curve (500 ng, 100 ng, 20 ng, 0 ng) and allowed to bind to the coated goat anti-mouse antibody. The ELISA finished using an anti-mouse IgG2a conjugated to HRP as a detection at various working dilutions selected based on manufacturer’s suggested dilution. When the curve was developed and read at 450 nm the acceptance criteria was met at the 1:5000 detection dilution using 2 μg/mL capture antibody coating (Figure 2D).

For IgG2c and IgA, the normal mouse antibody isotypes were coated directly to the high binding microtiter plate at standard concentrations (500 ng, 100 ng, 20 ng, 0 ng). Anti-IgG2c or anti-IgA detection antibodies conjugated to horseradish peroxidase were tested at various working conditions based on manufacturer’s suggested dilution. TMB was reacted then stopped with 2N H₂SO₄ and measured at 450 nm. Standard curves were interpolated and correlation coefficients (r²) using GraphPad Prism version 9.5.1 (Figures 2E and 2F) (See: Quantification and statistical analysis). Selection criteria for IgG2c and IgA were the same as for IgG and IgG1.

To assess the internal variability of the assay, standards were run under optimized assay conditions in replicates of six. Standards were run at concentrations to cover a range of the standard curve similar to standard optimization. The optical readouts for each dilution were averaged and the coefficient of variability calculated from the mean and standard deviation for each antibody isotype to assess the variability between replicates (Table 2). A coefficient of variability (%CV) of less than 10% is considered acceptable for intra-assay replicates.³ The standards for each isotype were below the 10% cutoff indicating a low degree of inconsistency within the assay. Total IgG had the greatest %CV of the five isotypes.

Table 2.

Intra-assay variability across plate replicates for each antibody isotype

Antibody isotype	N	Mean OD (SD)			Coefficient of variability (%CV)
		500 ng	100 ng	20 ng	500 ng	100 ng	20 ng
Total IgG	6	2.55562 (0.116)	1.6872 (0.109)	0.8283 (0.070)	4.52969	6.44848	8.41735
IgG1	6	3.81915 (0.048)	3.874517 (0.064)	3.122433 (0.038)	1.26724	1.676605	1.22582
IgG2a	6	2.3849 (0.040)	2.30925 (0.076)	2.061116 (0.036)	1.660375	3.275535	1.759253
IgG2c	6	2.1029 (0.036)	1.78955 (0.034)	0.88455 (0.031)	1.72098	1.913142	3.492162
IgA	6	2.107733 (0.023)	2.005933 (0.016)	1.544533 (0.013)	1.106593	0.786063	0.820133

Coefficients of Variability are calculated by dividing the standard deviation by the means then converted to a percentage. A coefficient of variability less than 10% indicates low variability between replicates.

To assess reproducibility between operators, three different end users independently repeated the intra-assay method outlined above. Results from each user were used to determine inter-assay coefficient of variability (Table 3). A coefficient of variability of less than 15% is considered acceptable for inter-assay replicates.³ The inter-assay values indicate an excellent degree of reproducibility of the protocol in the hands of different end users with %CV values less than 5% for each isotype across the range of standard dilutions.

Table 3.

Inter-assay variability across plates for each antibody isotype

Antibody isotype	N	Mean OD (SD)			Coefficient of variability (% CV)
		500 ng	100 ng	20 ng	500 ng	100 ng	20 ng
Total IgG	3	3.255344 (0.089245)	2.212306 (0.073795	0.8587 (0.042049)	2.741504	3.33566	4.89685
IgG1	3	3.890894 (0.069963)	3.918417 (0.053799)	3.0463 (0.054532)	1.798121	1.372966	1.790103
IgG2a	3	2.220489 (0.049606)	2.130094 (0.044768)	1.8592 (0.028741)	2.234024	2.101706	1.545861
IgG2c	3	1.983233 (0.057066)	1.154906 (0.049432)	0.459583 (0.015652)	2.877424	4.28019	3.40575
IgA	3	2.293606 (0.024891)	2.11975 (0.03234)	1.471439 (0.024869)	1.085227	1.525633	1.690109

Assays were performed by three different operators following this protocol. Coefficients of variability are calculated by dividing the mean standard deviation (N = 3) by the mean optical density (N = 3) then converting into a percentage. A coefficient of variability less than 15% on an inter-assay comparison indicates low variability between operators.

Quantification and statistical analysis

In an indirect ELISA protocol, the saturation binding of antibody to an antigen is measured to determine the ligand concentration that binds to half the receptor sites at equilibrium (K_d) and the maximum number of binding sites (B_max). Binding will not only be to antigen epitopes but may also bind nonspecifically to other parts of the antigen or other assay components. For accurate ELISA readings, it is important to exclusively measure specific binding. To do this nonspecific binding should be subtracted from the measured values.

Calculation of sample concentration is reliant on the interpolation of a standard curve. First, a table of specific binding is generated by subtracting the optical density value of nonspecific binding from the raw value of each standard and sample. The new optical densities for the standards can be analyzed using a nonlinear regression where the x-axis is the antibody concentration, and the y-axis is the value of specific binding (adjusted OD). The model for the linear regression is as follows:

$$Y=B_{\mathit{max}}\ast X/\left(K_d+X\right)$$

In this model, B_max is the maximum specific binding and K_d is the equilibrium dissociation constant. The dissociation constant is equal to the concentration of antibody required to achieve half the maximum binding at equilibrium. The concentrations of antibodies in the samples can be calculated with this formula. When using diluted samples, multiply the calculated concentration by the dilution factor to determine the concentration in the undiluted sample.

Many plate reader software programs can calculate these concentrations when set up properly before reading the plate. Otherwise, statistical software such as GraphPad Prism, R, or SAS can be used to interpolate a standard curve and values for the samples (Figure 3). Either of these options are recommended over manual calculation of standards and sample concentrations.

Figure 3.

Graphical representation of the standard curve mathematical model

If using GraphPad Prism the standard curve and concentrations of unknowns can be calculated by following these steps.

• 1.First, export the raw ELISA data to a text or Excel file.
• 2.Open Prism and create an XY data file. For the Y option select the second option and enter the number of replicates performed. If done in singlet, choose the first option under Y (Figure 4).
• 3.Once created you will see a spreadsheet window. In this spreadsheet enter the standard dilutions under the X values, note the 8^th value will be 0 to represent the blank. Under Group A list the replicate OD values for each standard in the same row as the standard value they correspond with. Below the standard OD values list the OD values of the unknown samples. Sample IDs can be listed in the row title if so desired (Figure 5).
• 4.Once the data is entered and sorted click on the “Interpolate a Standard Curve” button at the top of the page (under the Analysis section) to open the following option window. Select the appropriate model. For this procedure Hyperbola (X is concentration) is the desired model. Options can be selected or unselected here as desired (Figure 6).
• 5.After Prism completes the analysis a new page will be created with three tabs. The third tab provides test results including the correlation coefficient (r²) which serves as a control check for the data. The first tab, Interpolated X Mean Values, will list the concentrations calculated from the standard curve by sample ID. These mean values are the average of the replicate values from the second tab (Figure 7).
• 6.Interpolated values will be for the diluted serum. To calculate the concentration in the undiluted sample, multiply the interpolated value by the dilution factor used for the serum sample.
• 7.Since each ELISA plate has its own standard curve, repeat steps 3–7 for each ELISA plate by either creating a new data sheet or copying the previous family of sheets renaming the new sheet as necessary.

Figure 4.

GraphPad Prism start screen to begin data analysis, related to Quantification and statistical analysis step 2

Select an XY Table with enough Y replicates to match the number used for the samples.

Figure 5.

Data entry in XY datafile for ELISA analysis, related to Quantification and statistical analysis step 3

Standard curve concentrations are added to the X column with OD values for the concentration matched by row in Group A column. Unknown values are listed below the standards and sample ID for the unknowns can be listed in the row title square to the far left of the page.

Figure 6.

Statistical analysis option window in GraphPad Prism for ELISA analysis, related to Quantification and statistical analysis step 4

To interpolate unknowns from a standard curve, select “Hyperbola (X is concentration)”.

Figure 7.

Statistical analysis output from standard curve interpolation, related to step 4s

Model statistics can be found under the “Table of Results” tab. Interpolated results for unknowns are listed under “Interpolated X mean values” tab in the order they were entered into the datafile.

With concentrations calculated, analysis of the experimental data can be performed using appropriate statistical methods. For example, means and medians may be analyzed using parametric and non-parametric t-tests or ANOVA as appropriate.

Limitations

As with any ELISA, this protocol has its limitations. First, the length of the protocol can be time consuming requiring over one day when accounting for active time and incubation times. Second, ELISA readouts are often temporary, reliant on enzyme/substrate interactions which can degrade over time, therefore requiring readouts be obtained promptly once development is completed. Third, the indirect ELISA method is limited to a single antigen target. Measurement of different epitopes or cross reactivity to other targets requires modification to the protocol for those antigens.

Troubleshooting

Problem	Step	Potential cause	Potential solution
Edging effect (wells around side of plates have higher OD than middle)	23	Antibody binding varied by well due to temperature fluctuations across plate	•Seal plates •Do not stack plates during incubation, or do not stack more than 2 high to maximize temperature penetration of the plate.
Standard curve r2 below 0.98	23	Inaccurate antibody concentrations	• Remake standard curves paying attention to: ○ Pipetting steps ○ Volumetric calculations ○ pH of buffers
No color in wells (standards and samples)	23	Wrong substrate added to detection antibodies	•If stop solution has not been added, wash plates of substrate and repeat the development step (steps 19–21). •If stop solution has been added, repeat assay from the beginning.
		Reagents added in wrong order or not added at all	•Repeat following steps in order.
		Insufficient antibody	•Repeat increasing antibody concentration.
		Coating conjugate/antigen did not bind	•Use high binding ELISA plates
No color in wells (sample only)	23	Sample dilution too high	•Repeat assay with lower sample dilution
High background	23	Insufficient washing	•Increase number of washes or add a 30 s soak to each wash.
		Too much substrate	•Check substrate concentration and volume
		Insufficient blocking	•Check blocking buffer concentration and incubation time.
		Incubation times too long	•Ensure incubation times are not too long.
		DEA buffer pH	•Ensure DEA buffer pH is 9.8.

Problem 1

Higher absorbance in the outer wells of an ELISA plate (Figure 8) can occur when contaminations are introduced or when uneven temperatures across the plate occur (related to steps 2b. 4c, 4g, and 4k).

Figure 8.

Edging effect occurs when the outer wells of an ELISA have higher absorbance value than wells in the center of the plate or even their replicates

This may be caused by contamination of the outer wells or uneven temperature across the plate.

Potential solution

• •To prevent contamination, cover the plate(s) with a disposable plate seal or a clean reusable plate cover.
• •To ensure even temperatures across plates, stack the plates no more than 2 or 3 high when working with multiple plates.

Problem 2

If color is expected but no color develops (e.g., in standard wells) the culprit could be one of several (Figure 9). Using the wrong substrate (related to step 4n) can result in little to no color development. Adding the reagents in the wrong order will result in the antibodies not having targets to bind to (related to steps 4d, 4f, and 4h) resulting in no color development. Lastly, using reagents at dilutions too low to generate a strong signal can lead to no color development (related to steps 1b, 2a, 4d, 4f, and 4h).

Figure 9.

Lack of colorimetric development in any well may be the result of one or more process errors including using the wrong substrate, reagents not added or added in the wrong order, low antibody, or lack of coating

Potential solution

• •If the wrong substrate was suspected to have been added to the plates and stop solution has not been added to the assay, try washing the plate again and adding the correct substrate.
• •Repeat the assay paying attention to the order of reagent addition.
• •Check detection antibody is the appropriate pair for the standard (e.g., Anti-IgG2c on IgG2c standard).
• •Check antibody concentrations. Increase as necessary.
• •Use high binding ELISA plates.

Problem 3: Lack of color in sample wells (color development expected)

If color is expected in the sample wells but no color develops yet color developed in the standard wells (Figure 10) the culprit is likely the sample dilution factor. A sample dilution that is too high can dilute out a serological signal to a concentration too low to be detected (related to step 4d).

Figure 10.

Lack of colorimetric development in sample wells may be the result of low antibody concentrations in tested samples, if antibody presence is expected

Potential solution

• •Check sample dilutions. Repeat using a lower dilution to boost concentration to within standard curve.

Problem 4: High background

When working properly, the standard curve will return a decreasing optical density value correlating to the concentration of standard antibody with the blank wells having the lowest optical density values. If the blank wells have optical density values higher than expected (Figure 11) the reason is likely to be related to the wash step (related to steps 4a, 4e, 4i, and 4m), substrate concentration (related to step 4L), blocking buffer (related to step 3a), or incubation times (related to 4c, 4g, 4k, 4o).

Figure 11.

High background can have several causes including insufficient washing, too much substrate, insufficient blocking, or incubation times being too long

Potential solution

• •If the wash step is suspected to be the issue, increase the number of washes per step or add a 30 s soak to each wash step.
• •If the substrate concentration is suspected, check substrate concentration, expiration, and volume.
• •If blocking is suspected, check blocking buffer concentration and incubation time. Increasing either may help with background.
• •If incubation times are suspected, ensure incubation times are as close as possible to the stated times in the protocol.

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, David Dowling (David.Dowling@childrens.harvard.edu).

Technical contact

Further information on technical usage and application should be directed to and will be fulfilled by the technical contact, David Dowling (David.Dowling@childrens.harvard.edu).

Materials availability

This study did not generate new unique reagents. All non-experimental sample materials listed in this protocol are available from commercial vendors listed in the materials section.

Data and code availability

All data reported in this protocol will be shared by the lead contact upon request.

Acknowledgments

The current study was supported in part by US National Institutes of Health Development of Vaccines for the Treatment of Opioid Use Disorder Contract (75N93020C00038) to O.L. and D.J.D. The Precision Vaccines Program acknowledges the support of Drs. Wendy Chung, Nancy Andrews, and Kevin Churchwell of Boston Children’s Hospital. Figures 2, 3, 4, 5, 6, and 7 within this protocol were created with BioRender.com. Statistical figures were created with GraphPad Prism 9.

Author contributions

J.A.K. designed, performed, and analyzed the experiments and wrote the manuscript. A.P.-R., C.N.H., and E.B.N. assisted in the adaptation of the protocol. D.S. and D.J.D. oversaw the management of the project. K.W. and A.K. performed the inter-assay validation experiments. D.J.D. and O.L. conceived of the project. All authors provided feedback on the manuscript.

Declaration of interests

D.S., O.L., and D.J.D. are named inventors on vaccine adjuvant patents assigned to Boston Children’s Hospital. O.L. has served as a paid consultant to GlaxoSmithKline (GSK) and HilleVax. D.J.D. is on the scientific advisory board of EdJen BioTech and serves as a consultant with Merck Research Laboratories/Merck Sharp & Dohme Corp (a subsidiary of Merck & Co., Inc.). D.J.D., O.L., and C.N.H. are cofounders and sit as scientific advisors of Ovax, Inc. E.B.N. is a scientific advisor to Ovax, Inc. These commercial or financial relationships are unrelated to the current study. C.N.H. is named as an inventor on a patent for anti-fentanyl vaccine assigned to the University of Houston.