Case 10: Gadolinium (Gd) and Nephrogenic Systemic Fibrosis: Analytical Studies
Background
Nephrogenic Systemic Fibrosis (NSF), initially known as Nephrogenic Fibrosing Dermopathy (NFD), is a serious, potentially fatal, disease recognized in 1997 and first reported in 2000. NFD involved the skin but when systemic involvement of other organs was noted the name was changed to NSF. The link between this disease and exposure to Gd-based MRI contrast agents (GBMCA) was only recognized in 2006. The disease affects persons with severe Renal Insufficiency (mostly chronic, but sometimes acute), and is described in detail in the rapidly growing medical literature. The fibrosis in skin and other organs in NSF appears to be mediated by recruitment of bone marrow derived stem cells identified as "circulating fibrocytes" to the affected tissues.(See Pubmed for up-to-date links to the medical literature, and a website maintained by Dr. Shawn Cowper at Yale).
Gadolinium (Gd) (atomic number 64, atomic weight 157.2) is a member of the Lanthanide series, a Rare Earth Element (REE). There are a number of simple and complex organic molecules designed to chelate Gd to allow this paramagnetic element to be used clinically. Understanding the chemistry and relative stability of various GBMCA is essential to appreciating the relative risk for release of free Gd ions from the chelated agent, often thought to involve a process called transmetallation, in which another competing ion replaces the Gd bound in the chelate, thus releasing the free Gd ion. Ions such as Zn, Ca, Fe, Cu can each compete with Gd for binding to the chelate. Free Gd ions are known to be toxic to many cellular processes, e.g., acting as calcium channel blockers and killing macrophages (via apoptosis). The likelihood of release of free Gd increases as the GBMCA remains in the body for longer periods of time; as the clearance of these agents is nearly totally via excretion by the kidneys, the GBMCA clearance half-time is greatly lengthened in renal failure. However, it is noteworthy that even in persons with normal renal function, a small amount of the Gd from GBMCA is apparently released into the body and stored in the bones, most likely incorporated into hydroxyapatite. In a single Gd-enhanced MRI scan a dose (0.1 mmol/kg body weight) on the order of 1.5 grams of Gd is injected intravenously; of this approximately 1% of the Gd (15 mg) may be released and deposited in bones. [This calculation is based on (1) the reported 98-99% clearance of Gd, (2) the demonstration of several micrograms Gd/g bone measured following GBMCA administration, and (3) the assumption of total bone mass of approximately 10kg (see Abraham et al, Br J Derm, 2007)]
Gadolinium in Skin
Example of SEM backscattered electron imaging and EDS spectrum showing deposition of Gd in the dermis of this skin biopsy from a patient with NSF. Note the presence of P, Ca and Na also in the analyzed feature.
The Gd detected in these biopsies (days, months and years following the dose of GBMCA) represents insoluble Gd deposits, as the tissues have been in aqueous environment in vivo, followed by fixation of the biopsy in aqueous formaldehyde solution, and next dehydration processing through alcohols and xylenes, and then paraffin. Therefore, it is exceedingly unlikely that any intact GBMCA (which are water soluble) remains in the paraffin embedded tissue samples analyzed to date. Whether any intact GBMCA remains in the tissues many days, months or years after its administration is not resolvable by this technique, but it is unlikely that GBMCA remains intact in vivo for more than a few weeks. Recently, we reported that Gd deposits are not detectable in a series of skin biopsies from patients with normal renal function who have had GBMCA-enhanced MRI scans. [Boyd et al, J Am Acad Dermatol 59(2):356-8, 2008]
Analytical Methods
We developed and tested a quantitative method using Automated Feature Analysis in our SEM [Thakral and Abraham, J. Electron Microsc 56:181-187, 2007; doi: 10.1093/jmicro/dfm020]. This showed Gd deposits detectable in all the NSF cases analyzed to date (more than 60 biopsy and autopsy samples). This report also clearly showed that more Gd-containing features were detectable in the paraffin block analysis than in 5 µm paraffin sections mounted on carbon. This is a result of some penetration of the 20keV electron beam into the paraffin block to distances of greater than 5 µm. We also noted that the apparent diameter of Gd-containing features was less in the paraffin block than in the 5 µm section (from which paraffin had been removed). This is explained by the phenomenon of beam attenuation by the paraffin -- the deeper a feature is below the surface, the more the electron beam will be attenuated, making the detected feature diameter smaller than for the same feature located at the surface of the specimen.
Gd distribution in skin tissue sample.
Corresponding SIMS images of 40Ca (top) and 160Gd (bottom).
Our automated SEM/EDS method determines the number and composition of Gd-containing (and other) discrete deposits of insoluble inorganic materials in the tissues, reported in an area-based concentration (e.g., number of features and total x-ray counts per second per square millimeter of tissue cross section), and also reveals the spatial distribution of the detected features. This spatial distribution was found to be non-uniform and finding in some biopsies of much higher Gd concentrations several mm deep to the skin surface but not in the upper dermis should caution clinicians and other investigators that some biopsies may give a false non-detect or low concentration of Gd if not deep enough [see figure of distribution (modified from Thakral and Abraham, J. Electron Microsc, 2007)]. This method does NOT, and would not be expected to exactly match the gravimetric analysis by bulk analytical means such as ICP-MS. Each method has its strengths and limitations. Our automated SEM/EDS method is non-destructive and can note the associations and relative concentrations of multiple elements in single features, and the morphology and spatial distribution of a wide variety of features. It can detect a single feature as small as 0.25 µm diameter containing approximately 1 to 2 % Gd by weight in that feature. It cannot detect smaller features, nor will it detect Gd more diffusely distributed at concentrations less than approximately 1 to 2% by weight in the area probed by the electron beam. The ICP-MS method provides excellent and ëgold standardí gravimetric concentration but is destructive, cannot examine the associations of multiple elements in single features, nor the spatial distribution of the Gd within a given sample. That the SEM/EDS method cannot detect some of the insoluble Gd associated with Ca at lower concentration is shown using the more sensitive Secondary Ion Mass Spectrometry (SIMS) technique [Abraham et al, App. Surf. Sci (in press) 2008]. This representative figure shows the distribution of Ca matches to that of Gd for the majority of the tissue, with the exception of the one dark area in the Gd image (corresponding to a large multinucleated macrophage, approximately 50 µm diameter).
Further Research Needed
Investigation of NSF and its association with GBMCA is a rapidly evolving and fascinating subject, requiring input from a wide range of scientific disciplines. Some of the many questions which cannot be answered at present include: why some people receiving GBMCA develop NSF and others do not, the chemical and cellular processes involved in development of NSF, whether Gd-induced disease will eventually be detected in persons with normal renal function, and the long term consequences of growing body burden of Gd in millions of people around the world.
