Chronic excess exposure to manganese (Mn) in occupational settings can result in impairment similar to Parkinson’s disease¹. The accumulation of manganese can be detected by R1 mapping since Mn is paramagnetic. However, Mn-exposed workers, e.g. welders, usually receive chronic exposure to iron (Fe) at the same time. Though Fe is primarily studied for its ability to affect R2, it can also impact R1 and likely presents a confounding variable when assessing brain Mn accumulation. At such low concentration of Mn and Fe, further examination on a potential interaction of the two metals on R1 is needed. We therefore performed a phantom study by using various combinations of Mn and Fe levels to validate an empirical model which considers interaction of iron and manganese and its ability to affect correlation of R1 relaxation rate and Mn contents.

Phantom preparation

In total three groups of phantoms were prepared. Series A consisted of five tubes that contained various manganese concentrations from 0.002 to 0.012 mmol/L. Series B consisted of five tubes with iron concentrations ranging from 0.1 to 0.8 mmol/L. In series C-F, seven tubes that had different mixtures of manganese and iron solutions were prepared in order to determine a potential Mn-Fe interaction (Table 1). NiCl₂ and NaN₃ was added to all tubes as tissue mimic modifier and as an antiseptic, respectively.

MRI measurement

R1 measurements were performed on a 3T Signa HDx scanner (GE) with an 8-channel head coil at 26 ^oC by using a two-dimensional fast spin-echo inversion recovery sequence with the parameters: TR=2000ms, TE=8.6ms, and six inversion times of 75, 180, 350, 650, 1100, 1680 ms. Acquisition matrix was 128 by 128, field of view was 240, and slice thickness was 3 mm. The R1 rate of each pixel was obtained from these images by least-squares method in Matlab (Mathworks, Natick, MA). The equation used to calculate T1 (=1/R1) was :

$$S = S_{0}\times[1 - f\times e^{\frac{}-TI/T1}+e^{-TR/T1}]$$

where S=signal intensity from images, S₀=signal of proton density, TI= inversion time, T1=relaxation time, TR=repetition time, and f=inversion factor.

Modeling of Relaxivity

The following empirical equation was used to test the relationship between the concentrations of Mn, Fe and the R1 relaxation rates is as follows:

$$R1=R0+ r_{Mn} C_{Mn}+r_{Fe} C_{Fe}+ r_{MnFe} {(C_{Mn}+C_{Fe})}^{2}$$

Here R1 is the measured relaxation rate from the experiment. R0 is the relaxation rate in the absence of metals. C is the concentration of Mn or Fe in mmol/L. The relaxivities for Mn and Fe are r_Mn and r_Fe, respectively, representing a linear and independent contribution of each metal. The relaxivity value reflecting a potential interaction between Mn and Fe is r_Mn-Fe. The data was fit with and without this final term and the fits were compared.

The measured relaxation rates for all the mixtures of Mn and Fe are shown in Figure 1. All the R1 are within the range of 0.56-1.0 s^-1, which is in-line with literature². Without iron (Series A), R0 was found to be 0.56 s^-1, and R1 values increase linearly with Mn concentration, with a slope r_Mn of 7.7 s^-1(mmol/L)^-1. If iron is added (Series C and D), the R1 value is increased by roughly 0.3×C_Fe, with the slope of the curve unchanged. Fe presents a similar scenario as Mn: the relaxivity r_Fe was found to be 0.3 s^-1(mmol/L)^-1 in Series B, and the R1 of the curve increases by roughly 8×C_Mn (Series E and F), with C_Mn being the concentration of added Mn. The regression model gave a good fitting result (r²=0.974) even without including the cross term accounting for potential interaction. If we add the last term into the model, the fitting coefficient increases to r² = 0.993. The relaxivity of the cross term r_Mn-Fe is 0.97 s^-1(mmol/L)^-1, which shows a light but considerable contribution of an Mn-Fe interaction to the R1 relaxation rate.Our results are in line with previous studies showing that the relaxivity for Mn is around 7 s^-1(mmol/L)^-1. ³ The mathematical model from our results suggests that a model including only single linear terms for Mn and Fe already explains the R1 data with good accuracy. However, our result show that the fit is yet improved with the introduction of a cross term of Mn and Fe interaction into the equation. As a first step towards absolute quantification of brain Mn in Mn-exposed humans, our study suggests an interaction between Fe and Mn that contributes to the spin-lattice relaxation process. Future experiments will need to be performed in protein solutions that better resemble the interstitial brain milieu.