Covance is the drug development business of Laboratory Corporation of America® Holdings (LabCorp®). Content of this material was developed by scientists who at the time were affiliated with LabCorp Clinical Trials or Tandem Labs, now part of Covance.

Validation of the Meso Scale Discovery (MSD®) V-PLEX™ Cytokine Panel 1 (10-Plex) in Normal Human Serum (NHS)C.L. Evans, M. Saxena, K.E. Jeltema, S.E. Henson, A.E. Lopez and J.F. Marcelletti

AbstractPurpose: To validate the MSD® Cytokine Panel 1

V-PLEX™ assay system in NHS. This multiplex assay

allows for the simultaneous analysis of GM-CSF, IL-1α,

IL-5, IL-7, IL-12/IL-23p40, IL-15, IL-16, IL-17A, TNF-β and

VEGF. It has been validated following fi t-for-purpose

(FFP) principles by MSD®. We sought to corroborate

their FFP fi ndings and extend the analysis to include

method validation parameters using NHS as matrix.

Methods: MSD® V-PLEX™ Plus kits were performed

according to manufacturer’s recommendations. NHS

samples were screened for optimal minimum required

dilution and endogenous levels of the listed cytokines

were determined. A pool of NHS was used to generate

validation samples (VS) for the 10 cytokines. Multiple

LQC and LLOQ VS were evaluated to fi nely delineate

the lower limits of quantitation for the individual

analytes. Calibration standards (CS) and MSD®

controls (HQC, MQC, LQC) were assayed according

to instructions. Validation parameters in this on-going

study include accuracy and precision (A&P), selectivity,

specifi city, linearity on dilution, and stability.

Results: A 5-PL curve fi t for the CS yielded a mean

R2 > 0.9 for all cytokines. Lower limits of detection

(LLOD) of ≤ 1.1 pg/mL were observed for all cytokine

CS except IL-16 (LLOD ~ 2.7 pg/mL). Upper limits of

quantitation (ULOQ) ranged between 150-1500 pg/mL.

ULOQ was limited only by the highest concentrations

of the respective CS tested, since none of the cytokines

exhibited a plateau in ECLU. Inter- and intra-assay %CV

and % bias for back calculated CS were <20%. MSD®

buffer controls yielded similar %CV and % bias, except

for VEGF. Selectivity was established by spiking NHS

samples with the cytokines between their respective

LLOQ and LQC levels. Stability testing is underway

at the time of abstract submission. Typical detection

ranges for the 10 cytokines in NHS were documented

for the 10 assays.

Conclusions: We conclude that these results

1) corroborate the FFP fi ndings obtained by MSD®,

and 2) demonstrate the high likelihood that the MSD®

V-PLEX™ Cytokine Panel 1 assay system can be

adapted to fully validated status in NHS.

Introduction Methods ResultsThis work was designed to further validate the human Cytokine Panel 1 V-PLEX™ Plus biomarker assay system launched by Meso Scale Discovery (MSD®) using normal human serum (NHS) as sample matrix. The kit allows for the simultaneous assay of human specimens for GM-CSF, IL-1α, IL-5, IL-7, IL-12/IL-23p40 (IL-12/IL-23 hereafter), IL-15, IL-16, IL-17A, TNF-β and VEGF. The kit includes calibration standards (CS) and 3 levels of controls, H/M/L. The system has been very well characterized by MSD® using a fi t-for-purpose (FFP) approach. The validation included analyses for sensitivity, specifi city, accuracy and precision (A&P), matrix effects, stability and normal reference intervals for the cytokines. Many of the MSD® conclusions, such as lower limits of quantitation (LLOQ), are based on results with CS in buffer. The controls (high, medium, low) are prepared in non-human matrix rather than human serum or plasma. Therefore, gaps exist in validation parameters as specifi ed by guidance documents and white papers regarding full bioanalytical method validation (BMV) for drug pharmacokinetic (PK) ligand-binding assays (LBA) generally.

The FFP approach has been a widely accepted means to characterize biomarker assays in anticipation of preclinical and clinical trial sample analysis. The approach recognizes that not all aspects of BMV, as prescribed by guidance documents and industry standards, are necessarily feasible. Moreover, not all aspects of BMV are necessary depending on the intended use of the data derived. However, biomarkers are increasingly being used as drug co-diagnostics and clinical endpoints necessitating full BMV when possible. Signifi cantly, advances in preparation and characterization of biomarker assay reagents have made it possible to approach full validation of such assays in a manner consistent with drug PK bioanalytical methods.

However, full BMV of biomarker assays in matrix such as NHS is complicated by at least three important considerations.

■ First, biomarkers are often present endogenously in normal bodily fl uids and tissues.

■ Second, factors able to inhibit or augment the assay signal can also be present in such specimens.

■ Finally, individual biomarkers can be heterogeneous in the host and model systems using recombinant proteins may not accurately refl ect that heterogeneity.

Nevertheless, we’ve undertaken to fully validate this multiplex assay to the extent possible using VS prepared in NHS. The validation analytical plan includes the full range of parameters expected for a regulated, clinical study involving human serum samples, including A&P, selectivity, etc. This poster presents our progress to date in this endeavor.

TABLE 1CS Limits of Quantitation & Inter-assay

Reproducibility of MSD® Controls

Table 1. CS Limits of Quantitation & Inter-assay Reproducibility of MSD® controls. For Table 1, LLOQ is the lowest respective concentration of CS capable of reproducibly exhibiting a %CV <20% and %AR between 80-120%. The upper limit of quantitation (ULOQ) is the highest respective concentration of cytokine tested in the standard curves. The ULOQ values shown exhibited CV <20% and AR between 80-120%, respectively. MSD® Plus kit controls were tested according to manufacturer’s instructions in 3 independent runs by 2 analysts. Inter-assay statistics are shown. Only VEGF MSD® controls failed to meet full acceptance criteria, in particular that of %AR as shown in red.

FIGURE 2Cytokine Panel 1 Analytes Detected in NHS Specimens

Figure 2. Cytokine Panel 1 Analytes Detected in NHS Specimens. NHS samples were screened in the multiplex assay at the minimum required dilution of 1:4. Shown are mean ECLU obtained from duplicate determinations per sample. Also shown are the proportions of specimens (N = 20) that exhibited a signal above the LLOD. LLOD values (black symbol) were defi ned as the respective mean of the blank control values plus 3 standard deviations of that mean. LLOQ values are identifi ed as red symbols and were derived from CS curves as outlined for Table 1.

FIGURE 1Characteristics of Cytokine Panel 1

Calibration Standard (CS) Curves in Assay Diluent

Figure 1. Characterization of Calibration Standards (CS) in Assay Diluent. CS blend was reconstituted in Diluent 43 and inoculated into wells of the multiplex 96-well plate. Data are expressed as ECLU vs. nominal concentration (Panel A), and derived back calculated cytokine concentrations vs. nominal (Panel B). These latter data points were calculated by regression using a 5-PL curve fi tting model and Watson LIMS™ software.

Principle of the Method. The methods follow closely to that of the manufacturer, except as required to incorporate NHS matrix. In this method, the amounts of the ten listed biomarkers are measured by a sandwich immunoassay. The kit provides a 96-well plate pre-coated with capture antibodies on independent and well-defi ned spots. Each well of the 96-well MSD® plate is coated with ten specifi c antibodies that can each capture one of the ten desired biomarkers. Samples and a solution containing ten detection antibodies conjugated with SULFO-TAG™ are added over the course of sequential incubation and wash steps. Analytes in the sample bind to the capture antibodies immobilized on the working electrode surface (spots), and also bind the corresponding detection antibodies, completing the typical sandwich immunoassay confi guration. MSD® read buffer is added and the plate is loaded into the MSD® instrument where a voltage is applied to the plate electrodes causing the capture labels to emit light as electrochemiluminescence units (ECLU). The instrument measures the intensity of emitted ECLU to provide a quantitative measure of analytes in the sample. A regression of the ECLU of the standard curve samples versus concentrations is performed using Watson LIMS™ software and the concentrations of the analytes in the samples determined.

Preparation and assay of CS, VS, and MSD® controls. The lyophilized calibrator blend provided with the kit was reconstituted with MSD® Buffer (Diluent 43) and then serially diluted in Diluent 43 to obtain the desired concentrations for the standard curves. NHS samples were purchased from Bioreclamation and screened for suitability for use in a NHS matrix pool. Seven levels of VS were prepared by spiking this pool of NHS with the MSD® calibrator reference material. All VS contain at least 95% NHS. MSD® controls were reconstituted with Diluent 43. Diluted calibrators were added directly to the plate without further dilution. VS, MSD® controls, and NHS samples were diluted 1:4 in Diluent 43 for a fi nal minimum required dilution (MRD) of 25% when added to the wells. All determinations are assayed in duplicate (2 wells).

TABLE 2Verifi ed Limits of Quantitation Using NHS Validation Samples

Table 2. Verifi ed Limits of Quantitation Using NHS Validation Samples. VS were made at the concentrations shown using NHS as matrix. The VS were diluted 1:4 in assay diluent and assayed for cytokine levels. All VS were subjected to 6 accuracy and precision runs. The data presented identify those concentrations which passed acceptance criteria as indicated by blue boxes, and those that did not pass as represented by red boxes. It should be noted that the concentrations shown are before dilution 1:4, so the fi nal amounts in the assay wells were one-quarter of the illustrated values.

TABLE 3Validation Summary

MSD® V-PLEX™ Cytokine Panel 1 in NHS

Table 3. Validation Summary for the MSD® V-PLEX™ Cytokine Panel 1 in NHS. The results from this bioanalytical method validation are presented in Table 3 and summarized as follows:

1. The validation parameters were defi ned using VS made with ≥ 95% by content NHS pool.

2. The upper and lower limits of quantitation are based on accuracy and precision results of such VS against the buffer standard curve.

3. The respective LLOQ and ULOQ were defi ned by meeting acceptable intra- and inter-assay %CV and %AR. The assay ranges are suited for detection and quantitation of the studied cytokines in disease serum samples.

4. The calibration standard blend serially diluted in NHS exhibited highly signifi cant dilutional linearity for all cytokines tested (R2 ≥ 0.99).

5. The assays displayed acceptable selectivity for all cytokines except TNF-β and VEGF.

6. The cytokines exhibited the anticipated stability as spiked into NHS (not shown), with acceptable recovery observed with:

a. Minus 70°C long-term storage (LTS) for 132 days,

b. Four freeze-thaw (F/T) cycles, and

c. Five hours incubation at 4°C.

7. In conclusion, the MSD® V-PLEX™ Cytokine Panel 1 has been evaluated using NHS as sample matrix and all assays, except those for TNF-β and VEGF, are suitable for use in analysis of clinical serum samples.

Figure 3. Comparison of CS in Buffer vs. CS in NHS. CS blend was diluted in buffer and NHS pool. The high NHS CS solution was serially diluted 1:2 in matrix, before diluting 1:4 in assay diluent. Shown in Panel A are derived concentrations obtained using the respective cytokine standard curve in buffer. The values in Panel A depict total cytokines observed, endogenous plus exogenously added. Shown in Panel B are the values after subtracting baseline endogenous concentrations for the cytokines. These observations make the following points, including 1) human serum samples with high levels of these cytokines can likely be diluted if above ULOQ with valid concentration determinations (i.e., linear on dilution), and 2) the validity of using a buffer CS curve for quantitation of these cytokines in human serum.

FIGURE 3Comparison of CS in Buffer vs. CS in NHS