Critical Analysis Of The Paper On Magnetic Iron Oxide Nanoparticles

Main goal of the paper

To combat challenges that magnetic iron oxide nanoparticles face in medical applications, such as surface ligand functionalization, or in the process of metallic coating MIONs, where their iron oxide core could be exposed or magnetic properties changed, the authors were interested in investigating both the synthesis and performance of silica-gold coating process for a single nanoparticle system, designed for multi-modal imaging and heating. Specifically, they were curious regarding their synthesized nanoparticles’ ability to demonstrate magnetic heating properties consistent with original iron oxide core, with x-ray contrast, for imaging and laser heating applications. Thus, Woodard et al. aimed to perform a multi-step synthesis to coat Magnetic iron oxide nanoparticles MIONs with first a silica layer, then a gold layer to produce and characterize gold-coated iron oxide polycrystalline core and magnetic core-shell nanoparticles(iron oxide/silica/gold core-shell nanoparticles). To see if silica intercalated between iron oxide crystallites to form a dense core, the authors tested synthesized nanoparticles via small angle neutron scattering (SANS). They aimed to characterize NPs with TEM, DLS, magnetometry, SANS, T2-MRI, confocal microscopy, among other techniques. To test dual-modality imaging and MION detection concentrations, the authors aimed to perform MRI and x-ray CT contrast to characterize gold-silica-MIONs. To see if the original magnetic properties of the MIONs were altered, the authors tested magnetic characterization and heating with alternating magnetic fields. The authors also aimed to test heating performance with both magnetic and laser fields, with lastly supplementary in vivo imaging and heating of a mouse subcutaneous xenograft model of human prostate cancer.

The application

To synthesize gold colloid suspension, the authors added 600 mL of 1M aqueous sodium hydroxide and 2 mL of 1. 2 mM aqueous THPC to 90mL deionized water under 10 minute stirring, after which they added 3. 4 mL of 1%wt chloroauric acid. JHU MIONs were prepared by high-gravity controlled precipitation with thermal aging, and citric acid stabilization. To synthesize Au-SI-MIONs, the authors coated JHU MIONs with silica by consecutively adding JHU MIONs and 30% ammonium hydroxide to an aqueous ethanol solution. The authors then sonicated the nanoparticle mixture before then adding TEOS and left solution shaken overnight. Silica coated NP were centrifuged 3x with ethanol then re-dispersed in water, added APTMS, and the solution was shaken again overnight. The authors then seeded the resulting amino terminated silica surface with gold(1-2 nm) THPC colloid suspension and diluted THPC precursor solution with aqueous K2CO3 and sonicated. The authors added 1M Aqueous sodium chloride and amino-terminated nanoparticles to solution and sonication again and left overnight at 4C before centrifuging gold seeded nanoparticles with aqueous K2CO3 and 3x via permanent magnet and redistributing NP in aqueous K2CO3 with 1% HAuCl4 to coat the NP with gold. After vortex, 50% hydroxylamine addition turned the solution dark purple and resulting NP were washed again 3x with aqueous K2CO3 via permanent magnet. To characterize their synthesized nanoparticles, the authors performed DLS by measuring hydrodynamic diameter of uncoated Fe3O4/γ-Fe2O3 and coated Si or Au-Si nanoparticles on a Zetasizer Nano in 1. 8 mM K2CO3 with 0. 01 wt % nanoparticles. The authors specified a refractive index of 1. 33 for Fe3O4 and 2. 42 for DI H2O.

The authors then performed TEM with 10 μL of nanoparticle suspension in 100 μL of water dried for 24 hrs. on carbon coated copper grid at room temperature. The authors performed Magnetometry at 300 K from ± 5, 570 kA/m and at 5 K from ±5, 570 kA/m via SQUID VSM. The authors measured 5K data post cooling in 0 MF or in a 5. 6 MA/m MF. Next, characterization was performed with SANS, in which the authors used neutron wavelength of 0. 84nm in transmission for unpolarized SANS data. The authors used scattering vectors (Q) from 3e−5 to 5e−1/Å using 15m, 4m and 1. 33m detectors and measured their samples at 25C in H2O. For post-processing, the authors reduced raw 2D data obtained from the SANS experiments to obtain corrected 1D data.

The authors then corrected raw 2D data for empty cell, solvent, and background scattering as well as detector non-uniformity. A variety of geometrical functions fit the entire scattering curve per sample with SasView since complex NP structure needed linear combination of multiple model functions. The authors then performed T2-weighted magnetic resonance imaging, where they acquired T2-weighted images of gel phantoms with MIONs from 0 to 80 μg Fe/ml via horizontal bore spectrometer using the following spin-echo sequence parameters: repetition time (TR) of 4000 ms, echo time (TE) at 4, 8, 12, 16, 20, 24, 28 and 32 ms, slice thickness of 40mm, resolution of 128×128 pixels, and afterwards the authors analyzed reconstructed images with ImageJ.

To analyze MR MION contrast capabilities, the authors imaged phantoms with iron concentration from 0–80μg/ml and calculated NP R2 transverse relaxivity coefficient, for NP contrast efficiency as 155 (MION 1), 99 (MION 2) and 68mM−1 s−1 (MION 3). To compare the heating rates of JHU MIONs and Au-Si-MIONs, the authors delivered 5. 5W 780 nm laparoscopic laser excitation to both nanoparticle solutions and subsequently measured the temperature increases via FLIR and normalized SARs from iron content. The authors also performed X-ray computed tomography (CT) imaging on gel samples with 0–7 mg Fe/ml NP at 65 kV and 0. 7mA via SARRP, and reconstructed images with 1800 projections and used ImageJ to calculate Hounsfield units. Also, the last of the NP in vitro characterization was the AMF and SLP measurements estimated from heating data using mass of iron in the sample for iron oxide-based nanoparticles, heat capacity of the sample and, measured rate of temperature rise during heating interval. Then, to apply their synthesized nanoparticles in vivo, the authors used four male nude mice Athymic Nude-Foxn1nu strain, 5 -7 weeks old at 20 grams initial. The authors anaesthetized male nude mice presenting human prostate tumor xenografts and post pre-injection CT at 65 kV and 0. 7 mA via SARRP, the authors injected mice with JHU MIONs or AuSi-MIONs at ~5. 5 mg Fe/cm3 of tumor.

The authors designated a comparable control 20 uL saline intratumor injection. After injection of NP or saline, the authors obtained 2nd CT image reconstructed via ImageJ. The authors adjusted AMF system for stable oscillation at 150±5 kHz and 40 kAm−1 peak amplitude and assayed intratumor, rectal, and contralateral skin temperatures via fiber optic temperature probes. Afterwards, the authors performed histology on fixed tumors in 10Formalin and sectioned embedded paraffin blocks that were stained with H&E, Prussian blue, or silver enhancer and examined via Eclipse 80i microscope and obtained images at 4x mag and 20X objective.

Strong Points

The authors provided a helpful general cartoon schematic of their nanoparticle synthesis, illustrating in Figure 1 their process of silica coating, amino terminating, gold seeding, and finally gold plating their Au-Si MION NPs. Figure 2A has a clear illustration of size distribution via DLS, showing AuSi-MIONs to be 145nm as the largest MION form confirming the gold coating as well as an optimal size for potential in vivo nanoparticle delivery (less than 200 nm). In Figure 4B, the authors interestingly demonstrated that a gold coating may show complementary CT contrast to MRI via AuSi-MIONs concentration-dependent signal intensity with CT at ~7mg Fe/mL and corresponding ~360 Hounsfield units, while JHU-MION and Si-MION are ~sub-100 HU at the same concentration of 7 mg Fe/mL.

Weak Points

While the synthesis of the nanoparticles was described in detail, there was little to no detailed information in experimental on where the JHU MIONs were acquired or details on the process beyond the general “thermal aging” discussed in the results — this is unclear and could be concern for protocol repetition for results verification especially as the authors spend a good deal of time in the results discussing the JHU MIONs characterization. The way the authors report their results is slightly confusing as they basically re-state their experimental methods while also commenting on their data analysis. The authors were unable to model fit SANS data until data from multiple complementary techniques, such as TEM and DLS were combined to identify the combination of SANS geometric models appropriate for each particle construct, so there may be chance of introducing bias into the sample sets when combining data between techniques to find a fitting model, but it is unclear as they don’t specify what kind of data they pulled from or the techniques that they used if on the same data, very vague. This is important because it is unclear why the authors are making certain mathematical decisions to fit their data, seen as following ( almost like they are trying to mold a conclusion that supports their hypothesis): “the models were correlated between samples by using the parameters determined in the previous model. ” It is difficult to tell the distribution of the coating on the JHU MION cores in Figure 2c also with the addition of layers in 2 and 3, especially proof of concept for gold shell. Based on the various metal deposits in the tumor tissue samples in Figure 8 imaging, there may be potential heightened immune response or toxicity in surrounding tissue if used in vivo (especially since they note that the gold shell coating was degraded EXPOSING IRON OXIDE), but that would require further mandatory testing.

Possible “novel” future applications

As the authors were able to show sensitivity of magnetic resonance to detect low concentrations of the Au-Si-MIONs in tissue, as well as while the gold shell enabled x-ray visualization of higher nanoparticle concentrations, there is a versatility for the usage of these constructs in therapeutic imaging applications such as in cancer detection due to heightened contrast, or hard to detect disease in other tissue-related structures (ex. ) alternative imaging vs. spinal tap?).

My suggestions

Synthesis schematic with regards to the reagents and chemical synthesis involved as the actual experimental section was personally a little hard to follow in terms of if the NP were seeded with gold then coated again, or seeded once as well as all the reagents were used and what their context of application was for each step. Also, I want a more detailed protocol concerning the preparation of JHU MIONs, the first step of the NP synthesis. Also, I want to see the data asterisked as “not shown” for the Si-MION from Table 1, as it is incredibly important to show data with and without gold coating especially if used to test in vivo. I would like elucidation of the geometrical fitting that the authors briefly described when modelling their SANS data, such as how exactly they combined data from multiple techniques, like TEM and DLS, to be able to mathematically follow their logic to the same conclusions they are stating in Figure 3a. In Figure 7 I would want to see the tumor region again with added contrast. Even though histology confirmed the presence of gold and iron oxide deposits in Figure 8, I would want to do an in vivo immune assay to check for clearance as well as toxicity, since prolonged exposure to metals can be bad even if there is antitumor efficacy. Additionally I would want to study more mice to confirm the authors’ characterization of heating therapeutic potential.

Final decision

I would not fund this paper. I think the way they report their data is incredibly confusing, especially their decisions to highlight some irrelevant things while not including important data and extrapolating for conclusions…unclear how long it would take to get these constructs out of the system, or best for post mortem analysis.