Introduction & Background
Sanjiv (Sam) Gambhir (1) |
Last week in class, Dr. Islas
mentioned that there is an “art” to surgical oncology. Dr. Islas explained how surgeons will often
excise the tumor as well as additional healthy cells surrounding it, to ensure
they are removing the entire tumor and no residual cancer cells are left. Over the weekend, I came across a study
conducted by a group of Stanford scientist who have developed nanoparticles that
can home in on brain tumors, thereby increasing the accuracy of surgical
removal.
The senior author of the study is Sanjiv (Sam) Gambhir, MD,
PhD, professor and chair of radiology at the Stanford University School of
Medicine. The research was published
online on April 15, 2012 in Nature
Medicine.
Two common features of glioblastomas make them especially
difficult for neurosurgeons to remove. Glioblastomas
are characteristically described as rough-edged tumors, because small
“finger-like” projections often extend from the central mass and protrude into
healthy tissue along paths of blood vessels and nerve fibers (1). In addition to these “finger-like”
extensions, micrometastases further complicate the process of glioblastoma resection. Micrometastases are miniscule tumor patches
that have spread from the primary tumor to other sites (3). Scattered amongst healthy tissue,
micrometastases are invisible to the surgeon’s naked eye and if left unnoticed
they can readily grow into new tumors (1).
As Dr. Islas said when surgeons tackle tumors in the body they
can usually afford to remove some healthy tissue surround the cancerous mass in
hopes of getting rid of every harmful cell.
But, in the brain, each cell is important and too critical to sacrifice. As Dr. Gambhir says, “With brain tumors,
surgeons don’t have the luxury of removing large amounts of surrounding normal
brain tissue to be sure no cancer cells are left. You clearly have to leave as much of the
healthy brain intact as you possibly can.” (1)
In this study, Gambhir and his team developed minuscule
nanoparticles that highlighted brain tumors, precisely delineating their
boundaries and increasing the ease of the tumor’s complete removal (1). The nanoparticles, engineered by Gambhir, are
tiny gold balls coated with various layers of imaging reagents. When injected intravenously into mice with
brain tumors, the nanoparticles diffused out of damaged blood vessels feeing
the tumors and attached to cancerous cells therein. The nanoparticle’s gold core along with their
specialized coatings are visible to three distinct methods of imaging –
magnetic resonance imaging (MRI), photoacoustic imaging, and Raman
imaging. Using this tri-modality imaging
method, Gambhir obtained high resolution images of the tumors before and during
resection in tumor-bearing mice.
Capitalizing on the strengths of each imaging technique to guide each craniotomy,
Gambhir successfully removed the bulk of the glioblastomas, along with its
“finger-like” projections, and micrometastases.
Before delving into exactly how the scientists went about their study,
let’s take a step back and talk about the nanoparticles, MRI, photoacoustic
imaging, and Raman imaging.
The nanoparticles
Each nanoparticle is composed of a 60-nm gold core
surrounded by a thin Raman-active outer layer that is further protected by a
30-nm silica coating. Maleimide-DOTA-Gd3+
was then attached to thiol (SH) groups on the silica (5) (Fig. 1a). The diameter of each nanoparticle measures
less than five one-millionths of an inch, which is approximately one-sixteenth
the size of a human erythrocyte (1). Upon
intravenous injection, the nanoparticles work by capitalizing on the damage
caused by brain tumors to the blood-brain barrier. In doing so, they diffuse through the “leaky”
vessels surrounding the tumor, enter the extravascular space, and accumulate in
cells that comprise the tumor. In other,
non-cancerous regions of the brain where the blood-brain barrier is intact, the
nanoparticles are too big to cross and are therefore unable to accumulate in
healthy brain tissue (5). Unlike other
agents used for imaging that have a short blood half-life (i.e. Gadolinium for
MRIs), a single injection of Gambhir’s nanoparticles are retained in tumor
brain tissue long enough (~ seven days) to allow for preoperative planning and intraoperative
resection of the tumor (5).
MRI
Magnetic resonance imaging (MRI) uses radio waves and a
powerful magnet linked to a computer to generate images of areas inside the
body (4). Currently, MRIs are used to
give surgeons a macroscopic outline of the tumor before surgery (5). Use of
MRIs during surgery is possible, but
as stated earlier, requires repeated administration and high dosages of
contrast agents. In addition, surgeons
often experience discrepancies between preoperative
and intraoperative MRI scans. Before
surgery, MRIs are well-equipped to generate images that outline a tumor’s
boundaries, but given the rapidly-growing nature of brain cancers and the
constantly changing dynamics of the brain, preoperative
MRI’s are often not representative of the tumor during surgery. MRIs are
also limited in the spatial resolution they can provide (5).
Photoacoustic imaging
In photoacoustic imaging, red shifted light pulses are sent
into the body, which interact with image-enhancing nanoparticles. Excitation of the nanoparticles cause them to
heat slightly (without damaging nearby tissue) (6). Thermal expansion of the nanoparticles
creates pressure in the form of ultrasound waves. The sound waves are then recorded by an
ultrasound transducer that generates a three-dimensional image of the
nanoparticles distribution in the tissue.
In this study, the nanoparticle’s gold-core served as a good material
for absorbing light, and subsequently a good material at producing sound (6).
Raman imaging
Raman imaging generates images based
on a given material’s Raman spectrum.
The Raman spectrum is a signature pattern consisting of several
wavelengths unique to the specific material (1). A compete spectrum is acquired at each and
every pixel of the image and then analyzed to create colored images based on
the material’s composition and structure (7).
In this study, the nanoparticle’s silica and maleimide-DOTA-Gd3+
outer coating served to generate the initial Raman signal, while the
nanoparticle’s gold-core amplified the signal to be captured and recorded by a
microscope (1).
Histological validation of
nanoparticle sequestration by brain tumors
Gambhir and his team used
immunohistochemistry to determine their image-enhancing nanoparticles targeted
tumor tissue, and tumor tissue only (5).
The researchers implanted different types of human glioblastoma cells
(transduced with GFP) into the brains of mice.
Once the tumors were well established, they treated the mice with the
nanoparticles via the mice’s tail vein.
They then stained with antibodies against GFP and CD11b (microglia
marker antibody) to visualize tumor cells and glial cells, respectively
(5). Gambhir then took samples of tissue
at sites where tumor cells intersected with healthy cells. Using Raman imaging in conjunction with
scanning transmission electron microscopy (STEM) Gambhir confirmed the
nanoparticles accumulated within tumors, but not in healthy brain tissue (Fig.
2). In addition, Gambhir’s lab found
that the nanoparticles highlighted the glioblastoma’s “finger-like” protrusions
as well as isolated microscopic tumor foci (5).
The results indicate that the
nanoparticles are able to diffuse through the damaged blood-brain barrier
surrounding brain tumors and accumulate in cancer cells therein, without
needing a specific target mechanism.
Nanoparticles guide brain tumor
resection in vivo
Gambhir and his team then demonstrated
that their nanoparticles along with their three-pronged imaging proposal, could
indeed facilitate tumor resection in glioblastoma-bearing mice. To do this, the researchers once again
implanted different types of human glioblastoma cells (transduced with GFP)
into the brains of mice. Twenty-four
hours after injection of the image-enhancing nanoparticles into the mice’s tail
vein, they used all three imaging techniques to visualize the tumors the human
glioblastoma cells had spawned (5). The
researchers then placed the mice under anesthesia and performed craniotomies. Gambhir used visual inspection only to sequentially remove quarters of
the brain tumor (Fig. 3a). Following
each resection step, MRI, photoacoustic, and Raman imaging were performed. After the tumor resection seemed to be
complete by visual inspection, Gambhir obtained another set of images using all
three methods. MRI scans and
photoacoustic imaging indicated the tumor had been removed. But, Raman imaging showed several remnant micrometastases
(Fig. 3b). On further histological
analysis of the tumor (using antibody staining against GFP (tumor) and CD11b
(microglia)), Gambhir found tiny “finger-like” extensions of the tumor protruding
into the surrounding brain tissue (Fig. 3c).
A Raman image taken of this same section also confirmed the presence of
these microscopic cancerous growths, due to the retention of the nanoparticles
therein (5).
Overall, the results suggest that neither MRI nor
photoacoustic imaging alone can distinguish healthy from cancerous tissue at a
microscopic level to highlight every last bit of a tumor (1). Raman imaging, however, can. With its high sensitivity, Raman imaging was
able to flag residual microscopic tumor deposits that were otherwise not
visible to the naked eye, MRI, or photoacoustic imaging. Impart to its high accuracy, using real-time
Raman images, Gambhir was able to go back in and remove leftover cancer cells
that otherwise would have been left untreated.
The engineering of functionalized nanoparticles in
conjunction with a tri-modality imaging strategy has great promise for enabling
more accurate brain tumor imaging and resection. Gambhir and his team’s proposed strategy
entails 1) learning of the tumor’s general position and character before
surgery using MRI, 2) excise of the bulk of the tumor mass using photoacoustic
imaging, and 3) removal of once-invisible residual tumor deposits using Raman
imaging (Fig. 4). Ultimately, Gambhir’s
triple-modality-nanoparticle approach will help radiologists and neurosurgeons
better visualize the intricate outline of tumors before and during surgery,
thereby allowing for more accurate and successful brain tumor resection.
MRI
|
Photoacoustic Imaging
|
Raman Imaging
|
|
How it works?
|
Radiowaves and magnet to generate
image of body
|
Light heats up the nanoparticle’s
gold-core causing them to emit sound.
The sound is measured by an ultrasound transducer to generate
real-time images of the body
|
The nanoparticle’s silica & Gd
outer coating gives off a weak Raman signal that is amplified by the
nanoparticle’s gold-core. The
amplification is measured by a spectroscope to generate real-times images of
the body
|
Proposed clinical use
|
Preoperative
- initial detection of the tumor
- surgical planning
|
Intraoperative
- guiding resection of the bulk of
the tumor
|
Intraoperative
- guiding resection of residual
micrometastases
|
Strengths
|
Detectability
|
- High spatial resolution
- 3D imaging
- Deep tissue penetration
|
- High spatial resolution
- High sensitivity
|
Questions & Thoughts:
After
reading the paper, a few questions and thoughts came to mind:
1)
What
are nanoparticles? How do they differ
from the fluorescent-antibodies Gambhir used to locate tumor cells and
microglia?
Answer: There are four major advantages of
imaging using nanoparticles versus imaging using molecules:
a. Signaling power. In molecular imaging, molecules are injected into the body
and locate other molecules (i.e. molecules indicative of cancer). But, in order for those imaging-molecules to
be detected outside the body they need to produce a very large signal. Nanoparticles are very good at producing
signals because they are very good amplifiers.
Therefore, nanoparticles are able to generate signals that are much
stronger than your typical imaging-molecules.
b.
Versatility. Nanoparticles can be customized with different molecules
based on their proposed used. (For
example, in this study Gambhir used a gold-core along with silica and Gd outer
layers.)
c.
Multi-modality
competence. Nanoparticles can be made so that they produce a signal for
positron emission tomography (PET) scanning, computed tomography (CT), and even
a third modality such as single photon emission computed tomography (SPECT)
scanning. (For example, in this study
Gambhir created nanoparticles for MRI, photoacoustic imaging, and Raman imaging.) In molecular imaging, the multi-modality
feature is harder to create.
d.
Diagnostic
and therapeutic abilities. Nanoparticles themselves can serve as both a diagnostic tool
and a therapeutic agent. Gambhir
believes that if you are going to go through the trouble to create
nanoparticles that can home in on abnormal tumors just to get an image, then
that particle should be able to act as a diagnostic and therapeutic as
well. Therefore, nanoparticles could
essentially act as an imaging agent, a diagnostic, and a therapeutic. In molecular imaging, the molecules are
strictly used for identification. (6)
2) Can the nanoparticles
be used to light up other tumors elsewhere in the body?
Answer: Yes, the nanoparticles can be used to
highlight cancer cells in the colon.
In fact, in one of Gambhir’s original studies with
nanoparticles (prior to the one I blogged about) he engineered particles to go
into the bowel and detect colorectal cancer.
The use of colonoscopy has significantly lowered colon
cancer mortality rates, but this method of detection has its drawbacks. Dr. Gambhir explains, “Colonoscopy relies on
the human eye. So, this screening too,
while extremely useful, still misses many cancer lesions such as those that are
too tiny, obscure, or flat to be noticed.” (10)
Through his nanoparticles, Gambhir has created a way to catch
hard-to-see colon cancer cells early.
The nanoparticles used in the colon cancer study are very
similar to the ones I described in the brain cancer paper. They have a gold core, surrounded by a
Raman-active layer, which is further protected by a silica coating. The only difference is that the colon
cancer-nanoparticles are further functionalized with molecules that target
colon cancer cells, rather than the malemide-DOTA-Gd that functionalized the
brain cancer-nanoparticles. This makes sense,
because while the brain tumor-nanoparticles worked by diffusing through the
damaged blood-brain barrier (and did not have a specific target mechanism), the
colon cancer-nanoparticles obviously cannot do this. So, Gambhir needed to attach molecules, specific
to colorectal cancer cells, on the outside of the colon
cancer-nanoparticles. Using the colon
cancer-nanoparticle’s gold core to amplify the Raman signal, Gambhir
demonstrated that his nanoparticles latched onto colon cancer cells.
His findings have huge implications for improving detection
of colon cancer. Dr. Gambhir says,
“Right now, flat lesions might be entirely missed by the endoscopist who might
be doing screening for colorectal cancer, but by having the gold nanoparticles
light up where the cancer might be hiding in the bowel, now the hope is that
the endoscopist will be able to act on a lesion they would have otherwise
missed.” (10)
One can then imagine, that just by functionalizing the outer
layer of the nanoparticles with molecules specific to certain types of cancer,
the particles can hopefully be used to detect a wide variety of tumors.
3) Are the
nanoparticles toxic?
Answer: No,
the nanoparticles are nontoxic.
In a previous study, Gambhir showed that nanoparticles
similar to the ones used in this study (except without gadolinium, Gd3+)
were non-toxic in mice, suggesting that they will also behave well in humans
(11). Researchers administered
nanoparticles to two groups of mice, each containing 30 female and 30 male
animals. The dose was 1,000-fold larger
than would be required to get a clear signal from the nanoparticles (10).
The first group of mice received the nanoparticles
rectally. Gambhir then monitored the
mice at five different time points, ranging from five minutes to two
weeks. At each time point, Gambhir and his
team took a variety of measurements.
They followed the mices’ blood pressure, electrocardiograms, and
white-blood-cell counts. They examined
several tissues for increased expression of antioxidant enzymes or
pro-inflammatory signals, which would indicate physiological stress on the test
mices’ cells. Gambhir also stained
tissues with dyes that tag dying cells.
By the end of two weeks, none of the tests resulted in signs of stress
to any tissues. To determine where in
the body the nanoparticles had accumulated, Gambhir scanned the mices’ tissue
using electron microscopy. He found the
nanoparticles remained confined to the bowel, thereby posing no systemic threat
to other organs. The nanoparticles were
quickly excreted.
The second group of mice received the nanoparticles
intravenously. Once again, Gambhir and
his colleagues found no significant signs of inflammation or other signs of
toxicity. Scavenger cells present in organs
such as the spleen and liver quickly sequestered the nanoparticles.
With strong evidence that the nanoparticles are safe in
mice, Gambhir is now filing for FDA approval to proceed to human clinical
studies of the nanoparticles for diagnosis of cancer.
4)
Overall, Gambhir’s study seems very promising. I would have liked
to seen Dr. Gambhir monitor the mice for sometime after their craniotomies to
see whether or not the brain tumors recurred. If he had done this and found
that remission rates are low, then Gambhir’s findings – that image-enhancing
nanoparticles selectively targets all cancer
cells, including finger-like projections and micrometastases, and therefore aids in complete tumor
removal – become much more powerful. To relate this back to
my first blog, CD47 Antibody: A Promising
Breakthrough (8),
the scientists in this study monitored the mice for an additional four months
after administering the anti-CD47 antibody treatment (9).
Only after finding no recurrence of breast cancer were they able to
make the claim that their treatment may be effective at eliminating all cancer
cells, even cancer stem cells. In this study, Gambhir claims his
tri-modality-nanoparticle approach could enable more accurate surgical
resection of brain tumors, which can subsequently help to improve the current
poor prognosis for GBM. He should then compare recurrence or
mortality rates of tumor bearing-mice that undergo tumor resection guided by
nanoparticle-imaging with that of mice that undergo tumor resection guided by
visual inspection only. It could very well be that the nanoparticles do
not accumulate in all the
“finger-like” projections or micrometastases, and if this were the case then we
would expect to see no difference in the long-term surgical outcome between
nanoparticle-image treated mice and visually inspected mice.
5)
This
may be useful information for groups interested in doing their cancer project
on new, cutting-edge cancer therapies.
Thanks for
reading!