br Fig Schematic workflow of
Fig. 4. Schematic workflow of: (A) the immune subtraction of transferrin using iodinated antibody coupled to AffiAmino Ultrarapid Agarose Magnetic beads and (B) the sandwich immunoassay coupled to ICP-MS detection of 127I for the determination of transferrin.
magnetically separate. Finally, the immune complex was disrupted with 0.1% HNO3 and introduced in the ICP-MS by flow injection analysis for monitoring iodine, as described in the procedures section.
First the analytical performance of the developed sandwich im-munoassay with ICP-MS detection (monitoring 127I) was evaluated by analysing two transferrin standards of different concentrations (10.3 and 16.2 µg mL−1). The analysis was performed in triplicate for each trans-ferrin concentration to determine the reproducibility of the metho-dology. For ICP-MS determination of iodine, a calibration curve was obtained by FI- ICP-MS using NaI standards of known concentration (from 25 to 500 ng mL−1 as I). This calibration curve is shown in Fig. 5. As can be observed, a good linear correlation (R2 = 1) was obtained that permitted the quantification of the two transferrin standards based on the measurement of iodine and on the iodine:antibody stoichiometry of 27:1 stablished in the first section. The amount of transferrin found in
SD), respectively. These results revealed transferrin recoveries of 88%
and 86% respectively, which suggests that the potential carbon en-hancement effect on the ICP-MS signal of 127I due to the degraded im-
mune complex in the plasma is negligible in this case. Finally, from the calibration graph of Fig. 5 the limit of detection (LOD) of the proposed methodology (3σ criterion) was estimated to be 140 ng mL−1 of trans-ferrin. This LOD is better than that obtained by HPLC-ICP-MS (502 ng mL−1) . Therefore, the proposed analytical strategy should be sui-table for the determination of transferrin in cell lysates. In this regard, two different breast cancer cell lines from two phenotypes of different malignancy have been used in these experiments: MCF-7 (minimally invasive) and MDA-MB-231 (highly invasive). Cell cultures and lyses were conducted as previously described (see Fig. 1).
The obtained results for the two different cell lines can be sum-marized in Table 1. The cell number of each of the cell culture analyzed is shown in the first column. However, since the cell number is a re-latively inaccurate parameter (estimated by hemocytometry) as re-vealed in a previous publication  we have used the wet DZNep weight
Fig. 5. Calibration graph obtained by flow-injection (FI)-DF-ICP-MS (mon-itoring 127I) using standard solutions of INa with an iodine concentration from 25 to 500 ng mL−1. The error bars (standard deviation) obtained at each con-centration for three repeated measurements are also shown. The inset shows the 127I signals obtained by (FI)-DF-ICPMS for triplicate measurements.
for normalization. Therefore, the concentration of transferrin is ex-pressed as µg of protein per g of cells. As can be observed in Table 1, the mean concentration of transferrin in the MCF-7 cell line is 2.3 ± 0.6 µg g−1 (mean ± SD, n = 6 independent cell cultures), while in the MDA-MB-231 cell line is 0.6 ± 0.2 µg g−1 (mean ± SD, n = 5 independent cell cultures). These two values can be considered statis-tically significant different (Welch's t-test), so the cell line MCF-7 (minimally invasive) shows approximately 3.5 times higher transferrin concentration than the MDA-MB-231 line (highly invasive). This finding is in agreement with the higher intracellular iron concentration detected in the same cell line in previous experiments (2-fold superior
Obtained results for the analysis of transferrin (Tf) in cell cultures based on Iodine determination by ICP-MS.
in the case of the MCF-7 line)  since transferrin is expected to be the main Fe transporter into cells. However, such higher Fe and transferrin concentration do not reflect the malignancy potential of the specific phenotype since the MDA-MB-231 is considered the most malignant model (more invasive). Previous studies conducted in this type of breast cancer cells have revealed that iron might play a dual function in tu-morigenesis . In one hand, an excess of Fe in normal cells might induce a neoplastic transformation via oxidative stress . On the other hand, in tumour cells, the dysregulation of extracellular and in-tracellular mechanisms necessary to maintain iron homeostasis might accelerate tumour growth and more aggressive tumour behaviour re-sulting in metastatic and recurrent processes [32,35].
Although current literature reveals a higher expression of trans-ferrin receptors in breast cancer cells, very little work has been done in the determination of transferrin cellular content. Thus, this works shows the suitability of the proposed strategy for the determination of this important protein, which is a key player in cellular iron home-ostasis, with high sensitivity and in extremely low sample volumes being an interesting alternative for total transferrin determination.