Matrix Effect Of PEG Precipitation In Detection Of Macroprolactin In IMMULITE^® AND IMMULITE^® 2000 Prolactin Assays

Abstract

We also performed studies on 40 samples that have been shown to contain macroprolactin by gel filtration or PEG procedures in the Delfia prolactin assay. Based on measurements with treated calibrators, the recoveries on these samples after PEG precipitation ranged from 6 to 67% (mean = 34%). For all samples (with or without macroprolactin) after PEG precipitation, when the prolactin values that were derived from the original calibration curve (Y) were compared to the values derived from the curve with treated calibrators (X), a linear relationship was found: Y = 1.38X + 0.31 ng/mL, r² = 0.9999. We conclude that in order to use a PEG precipitation procedure for evaluation of macroprolactin in patient samples, matrix effects should be taken into consideration, especially if there is an interest in defining the concentration of active prolactin in the macroprolactin samples.

Introduction
Macroprolactin is one of the multiple forms of prolactin in vivo. It is characterized as high molecular weight complexes of prolactin and IgG. Although it has little or no bioactivity, macroprolactin can be recognized in immunoassays with varied crossreactivities depending on the samples or the method used (John et al., 2000; Fahie-Wilson, 2000). This crossreactivity, combined with the fact that macroprolactin also has a longer half-life in vivo, may result in a false impression of hyperprolactinemia that could lead to diagnostic confusion or inappropriate patient treatment. It was reported that macroprolactin was the cause of 16 percent of the hyperprolactinemia results measured by the Roche Elecsys assay (Fahie-Wilson et al., 2000). A number of procedures have been described for identification of the presence of macroprolactin including gel-filtration chromatography and PEG precipitation. Because it is difficult to routinely perform gel-filtration chromatography, PEG precipitation has become the favored method.

Materials and Methods
IMMULITE and IMMULITE 2000. The IMMULITE is a random-access, solid-phase immunoassay analyzer that employs chemiluminescent enzyme detection and an efficient centrifugal wash. The IMMULITE 2000 employs an equivalent technology with additional features for higher volume immunoassay processing. (Diagnostic Products Corp., Los Angeles, CA.)

PEG precipitation. One volume of 25% polyethylene glycol (PEG) 8000 (Sigma, St. Louis, MO) solution in normal saline was mixed with 1 volume of patient sera, controls, adjustors, or calibrators. The mixture was allowed to stand at room temperature for 10 minutes, and then centrifuged at 3,000 x g for 15 minutes. Supernatants and untreated samples were assayed for prolactin with IMMULITE or IMMULITE 2000 automated chemiluminescent immunoassays. A precipitation method using PEG 6000 was also performed on samples assayed with the Delfia fluoroimmunoassay (Auto Delfia, Wallac, UK).

Gel filtration. Gel-filtration chromatography was performed on Sephacryl
S-300 (Pharmacia, Peapack, NJ). Macroprolactin was identified as a peak of immunoreactive prolactin eluting between IgA and IgG. If macroprolactin was detected, all fractions were analyzed to confirm the presence of two peaks of prolactin immunoreactivity, and macroprolactin was quantitated from the peak area by fluoro-immunoassay (Auto Delfia, Wallac, UK).

Results
Prolactin recovery after PEG precipitation
¹²⁵I-labeled monomeric prolactin tracer (25 µL) was added to (475 µL) each of 10 sera (475 µL) with no apparent macroprolactin. Next, either PEG diluent (saline control) or PEG solution was added. After precipitation, each supernatant (410 µL) was counted. Mean recovery was about 88% of values prior to precipitation.

Human IgG recovery after PEG precipitation
¹²⁵I-labeled human IgG tracer (60 µL) was added to each of 10 sera (1140 µL). Next, either PEG diluent (saline control) or PEG solution was added. After precipitation, each supernatant (400 µL) was counted. An average of 7% of added counts remained in the supernatants, indicating that 93% of the IgG was precipitated by the PEG procedure.

The results that follow are for the IMMULITE system. Similar results were obtained for the IMMULITE 2000 system. (Data not shown.)

Matrix Effect of PEG Precipitation on IMMULITE Assay
Random patient sera and controls (n = 16) with and without PEG precipitation were analyzed by the IMMULITE Prolactin assay. After correction for a dilution factor of 2, recoveries for the PEG-treated samples were calculated. An apparent over-recovery of 116 to 188% (143% ± 18%, mean ± SD) was observed. Calibrators included in the run were also treated by the same PEG procedure. All measurements were recalculated using the calibration curve that was generated from the PEG-treated calibrators, and a satisfactory recovery (100% ± 12%, mean ± SD) was obtained. This difference in values for samples with and without PEG treatment clearly demonstrated a matrix effect. Good recovery could be obtained when both calibrators and samples were treated by the same procedure (Table 1).

Table 1. Over-recovery of monomeric prolactin by the IMMULITE assay after PEG precipitation, and correction using PEG-treated calibrators.

PEG precipitation was also performed on 40 additional samples containing macroprolactin. For all samples after PEG precipitation, IMMULITE prolactin values derived from the original calibration curve (Y) and from the "treated" curve (X) demonstrated a linear relationship (Figure 1).

Figure 1. Comparison of IMMULITE Prolactin values after PEG precipitation, based on PEG-treated and untreated calibrators.

Y = 1.38X + 0.31 ng/mL   r2 = 0.9999

Y = 1.38X + 0.31 ng/mL   n = 56

Macroprolactin recovery by IMMULITE after PEG precipitation
We studied the 40 samples that had been shown to contain macroprolactin by PEG precipitation, gel-filtration chromatography or protein G chromatography in the Delfia prolactin assay. Prolactin levels by the IMMULITE Prolactin assay without PEG treatment were 12 to 120 ng/mL, with a mean of 42 ng/mL. Recoveries after PEG precipitation ranged from 8 to 95% (45% ± 24%, mean ± SD) by IMMULITE. Recoveries on these samples after PEG precipitation, based on treated calibrators, ranged from 6 to 67% (34% ± 17%, mean ± SD). (See Figure 2.) For normal samples that are not expected to contain macroprolactin (1 – 16), no data on the Delfia prolactin assay are available. For the macroprolactin samples (17 – 56), the percentage of monomeric prolactin was estimated in the Delfia prolactin assay by PEG precipitation for samples 17 – 37, and by gel-filtration chromatography for samples 38 – 56. Samples 47 – 56 were examined for the presence of IgG in the macroprolactin by affinity chromatography with protein G: a major component was detected in all samples except sample 53.

Figure 2. Percent recovery of prolactin with PEG procedure. Samples are grouped according to the method used to estimate the percentage of monomeric prolactin by the Delfia prolactin assay. (See text for details.)

Conclusion and Discussion
We have shown that the PEG procedure used in this study allows an 88% recovery of monomeric prolactin as verified by the ¹²⁵I labeling method. The efficiency of the procedure for precipitation of human IgG is about 93%, as verified also by the ¹²⁵I labeling method. Essentially, the PEG procedure allows satisfactory recovery of monomeric prolactin and effective precipitation of human IgG molecules.

The PEG procedure allows effective identification of samples with and without macroprolactin in IMMULITE and IMMULITE 2000 prolactin assays. However, a matrix effect cannot be disregarded, because it causes a significant over-recovery if not corrected for by using calibrators treated with the same PEG procedure. To use a PEG precipitation procedure for evaluation of macroprolactin in patient samples, matrix effects should be taken into consideration, especially if there is an interest in defining the concentration of active prolactin in the macroprolactin samples.

References
1. John R, et al. Clin Chem 2000;46:884-5.
2. Fahie-Wilson MN. Clin Chem 2000;46:2022-3.
3. Fahie-Wilson MN, et al. Clin Chem 2000;46:1993-5.
4. Gibson G, et al. Clin Chem 2001;47:331-3.
5. Urgell, et al. [poster PO-L017]. Clin Chem Lab Med 2001;39(Suppl):S244.
6. Berlitz, et al. [poster PO-L002]. Clin Chem Lab Med 2001;39(Suppl):S240.