Introduction
Endoscopic mucosal resection (EMR) is widely accepted as the standard treatment for
early gastrointestinal cancer [1]. Despite the convenience of EMR, tumor recurrence after EMR has been a problem because
of piecemeal resection of larger lesions. The technique of endoscopic submucosal dissection
(ESD), which enables en bloc resection of a larger lesion without compromising curability,
has rapidly become popular [2]. The creation of a submucosal cushion by fluid injection is essential for easy and
safe EMR or ESD. It is accepted that the “ideal” solution for submucosal injection
should provide a thick submucosal fluid cushion, remain in the submucosal space long
enough to safely allow EMR or ESD, and preserve tissue specimens to allow for precise
pathologic staging. Generally, viscous and hypertonic solutions tend to produce adequate
submucosal cushions, and a greater solution density provides a longer-lasting submucosal
cushion [3]; however, highly viscous solutions are difficult to inject through a catheter. There
is a demand for highly viscous submucosal injection materials retaining good injectability
through a catheter. Of the various solutions proposed for submucosal injection, sodium
hyaluronate (SH) has been reported to be the most effective. However, SH solution
is quite expensive and research on the development of new submucosal injection materials
is ongoing [3]
[4]
[5].
We have focused on a dispersion of cellulose nanofibers (CNF) as a potential new submucosal
injection material. CNF is prepared by making extremely fine naturally derived cellulose
fibers using nanotechnology. Typical fibril widths are 5–20 nm with a wide range of
lengths. CNF is a sustainable and renewable nanomaterial with many desirable features,
including high strength, high modulus, low coefficient of thermal expansion, and high
transparency. These basic properties of CNF make it an interesting material for many
potential applications in medicine, such as a covering material for skin wounds or
as an artificial vascular graft [6]
[7].
We focused particularly on the TEMPO-oxidized cellulose single nanofiber (TOCN) dispersion.
TEMPO (2,2,6,6-tetramethylpiperidin-1-oxyl) is a chemical compound which is used as
a mediator in controlled radical polymerization, and has applications in biochemistry
[8]. TOCN dispersion has a high gel-like viscosity in static conditions because of its
strong thickening effect and also exhibits high thixotropy, i. e., it becomes less
viscous when shaken or agitated, then returns to a gel or viscous state immediately
upon cessation of movement. Thus, thixotropy is a reversible behavior [9].
We hypothesized that TOCN dispersion could act as an effective submucosal injection
solution because of its high viscosity and thixotropy. The high viscosity of TOCN
dispersion should provide a long-lasting thick submucosal cushion, and its high thixotropy
could reduce its viscosity when it is injected through a catheter. The consequent
decreased injection resistance would allow the operator to inject more easily during
endoscopic treatment. In this study, we examined the catheter injectability and mucosa-elevating
capacity of TOCN dispersion to evaluate the feasibility of using this material for
submucosal injection.
Material and methods
Materials
The CNF concentrations of TOCN dispersions in this study were 0.35–0.45 %. CNF dispersion
takes a transparent gel-like form under static conditions and will become less viscous
and flow over time when shaken or agitated. When movement ceases, it will return to
a gel-like state almost instantly because of its high thixotropy. For analysis, 0.4 %
SH solution (Muco Up; Boston Scientific Corp., Natick, Massachusetts, United States)
and physiological saline (isotonic sodium chloride solution; Otsuka Pharmaceutical
Factory Corp., Tokushima, Japan) were used as the control and reference materials
to examine the catheter injectability and mucosa-elevating capacity of the CNF dispersion.
Subjective assessment of injectability of CNF dispersion into catheter
First, we analyzed the viscosity of physiological saline, 0.4 % SH solution, and 0.35–0.45 %
CNF dispersions under static conditions. Under conditions equivalent to those clinically
used in an endoscopic treatment, a 5-mL injection syringe (Terumo Corp., Tokyo, Japan)
with a 25-gauge endoscopic injection needle (Top Endoscopic Injection Needle Super-Grip;
Top Corp., Tokyo, Japan) was used. Ten adult men and 10 women were asked to inject
each solution into a free catheter. The patients were asked to rate how difficult
it was to perform catheter injection of each solution using a numerical rating scale
(NRS), with a higher score indicating more difficulty in injection. These patients
were blinded as to whether they were injecting 0.35–0.45 % CNF dispersions or the
0.4 % SH solution.
Objective assessment of the mucosa-elevating capacity of CNF dispersion in porcine
stomach
Porcine stomachs were divided into pieces approximately 3 × 3 cm, and stretched flat
on a corkboard with pins. Using a 2.5-mL syringe (Terumo Corp) with a 23-guage needle
(Terumo Corp), 2 mL of each solution was injected into the submucosa through the resected
margin. Mucosal elevation was observed and recorded with a digital camera (CAMEDIA
E-20; Olympus Corp., Tokyo, Japan) immediately and 5, 10, 20, 30, 60, 90, and 120
minutes after the injection. The height of mucosal elevation was measured from the
recorded images, and was defined as the distance from the mucosal surface before injection
to the top of the elevation. This experiment was repeated six times. We compared the
mean height of mucosal elevation produced by each solution.
The stomachs injected with each solution were fixed in 10 % neutral buffered formaldehyde
solution and embedded in paraffin. Each paraffin block was sectioned and stained with
hematoxylin and eosin for histological examination to evaluate the submucosal cushion
and the effect of the injection on the tissue.
Statistical analysis
We summarized the data using descriptive statistics. We used analysis of variance
with Bonferroni’s multiple comparison tests for post hoc analysis in Stata v.13.1
(Stata Corp., College Station, Texas, United States).
Results
Catheter injectability of CNF dispersion
[Fig. 1] shows the static viscosity of physiological saline, 0.4 % SH solution, and 0.35–0.45 %
CNF dispersions. The 0.35–0.45 % CNF dispersions had much higher viscosity than 0.4 %
SH solution. The viscosity of each CNF dispersion was directly proportional to the
CNF concentration. [Fig. 2] shows the results of the catheter injectability test. Catheter injectability revealed
resistance depending on the concentration of the CNF dispersion. Compared with the
0.4 % SH solution, 0.35 % and 0.4 % CNF dispersions showed no difficulty in catheter
injectability, whereas the injectability of 0.45 % CNF dispersion was clearly inferior
to that of 0.4 % SH solution with a statistically significant difference (P < 0.001). The 0.4 % CNF dispersion had the highest CNF concentration among the CNF
dispersions that had shown no difficulty in catheter injectability compared with 0.4 %
SH solution.
Fig. 1 Viscosity of physiological saline, 0.4 % sodium hyaluronate (SH), and 0.35–0.45 %
cellulose nanofiber (CNF) in static conditions. The viscosity of 0.4 % SH solution
was 67.8 mPa · s. The viscosities of 0.35 %, 0.4 %, and 0.45 % CNF dispersions were
16 000, 19 800, and 23 200 mPa · s, respectively. The 0.4 % CNF dispersion has about
300 times higher viscosity than 0.4 % SH solution.
Fig. 2 Catheter injectability of 0.35–0.45 % cellulose nanofiber (CNF) dispersion compared
with 0.4 % sodium hyaluronate (SH) solution. The average numeral rating scale (NRS)
score for the 0.4 % SH solution was 61.5 ± 11.0. The average NRS score for the 0.35 %
and 0.4 % CNF dispersions were 62.5 ± 9.2 and 66.0 ± 9.66, respectively, and there
were no statistically significant differences compared with that for the 0.4 % SH
solution (P = 0.35 and 0.11, respectively). On the other hand, the average NRS score for the
0.45 % CNF dispersion was significantly higher at 75.5 ± 7.97 than that for the 0.4 %
SH solution (P < 0.001). Furthermore, the NRS score for the 0.4 % SH solution was significantly
higher than that for physiological saline (P < 0.001). Individuals were asked to rate the difficulty of catheter injection of
the solution using the NRS score. The results are shown as mean and standard deviation.
N.S., not significant.
Comparison of the ability of submucosal injection solutions to maintain mucosal elevation
in porcine stomach
Based on the results of catheter injectability, the mucosa-elevating capacity in porcine
stomach was examined with 0.4 % CNF dispersion that had shown no difficulty in catheter
injectability and had the highest viscosity. The time-dependent changes in mucosal
elevation created by each solution are shown in [Fig. 3]. The 0.4 % CNF dispersion produced a significantly higher mucosal elevation than
the 0.4 % SH solution at all time points after injection.
Fig. 3 Chronological changes of mucosal elevation after submucosal injection of physiological
saline, 0.4 % sodium hyaluronate (SH) and 0.4 % cellulose nanofiber (CNF). The 0.4 %
CNF dispersion maintained a higher mucosal elevation compared to 0.4 % SH solution.
There was a statistically significant difference in the height of mucosal elevation
between the 0.4 % CNF dispersion and 0.4 % SH solution at all points (**P < 0.001). 0.4 % SH solution produced a higher mucosal elevation than physiological
saline (*P < 0.001).
[Fig. 4] shows typical histological images of the mucosal elevation induced by injection
of 0.4 % CNF dispersion. We noted accumulation of 0.4 % CNF dispersion in the submucosa
that separates the mucosa and the muscle. The mucosal and muscle layers were clearly
separated by submucosal accumulation of 0.4 % CNF dispersion, even 60 minutes after
the submucosal injection. The 0.4 % CNF dispersion injected into the submucosal layer
caused no apparent tissue damage as seen by microscopy.
Fig. 4 Histological examination of submucosal elevation immediately and 60 minutes after
injection of 0.4 % cellulose nanofiber (CNF) dispersion in gastric specimens (hematoxylin-eosin
staining, loupe image). The mucosal layer and proper muscle layer are widely separated
by the submucosal accumulation of 0.4 % CNF dispersion. Sufficient accumulation of
0.4 % CNF dispersion is observed in the submucosa even 60 minutes after the submucosal
injection.
Discussion
The results of our study demonstrated that 0.4 % CNF dispersion can maintain submucosal
elevation for a longer period than 0.4 % SH solution. Although the viscosity of 0.4 %
CNF
dispersion is about 300 times higher than that of 0.4 % SH solution under static conditions,
the catheter injectability of 0.4 % CNF dispersion was equivalent to that of 0.4 %
SH solution. The shear stress occurring by shaking or agitating changes the viscosity
of the CNF dispersion. When the operator begins to inject 0.4 % CNF dispersion, it
becomes less viscous and flows smoothly because of its high thixotropy index. Thus,
the high thixotropy of the CNF dispersion can effectively reduce the resistance to
injection and make it straightforward for the operator to inject it through a catheter.
Once the injected CNF dispersion reached the submucosal layer, it returned to its
original high viscosity almost instantly. This recovery of viscosity may have contributed
to its ability to maintain submucosal elevation for a longer period compared with
SH solution. The viscosity change of the CNF dispersion due to the change in shear
stress was stable and consistent ([Supplementary Fig. 1s]).
Supplementary Fig. 1s The viscosity change of 0.4 % CNF dispersion when shear stress is changed repeatedly.
Green dots were measured viscosity values when a fixed high shear rate was applied.
Blue dots were measured viscosity values when a fixed low shear rate was applied.
High shear stress reduced the viscosity of CNF dispersion rapidly and the decreased
viscosity value was constant. The viscosity value of the CNF dispersion decreased
to about 1/300 compared to before applying shear stress because of its high thixotropy.
Then, when the shear stress was decreased, the viscosity of the CNF dispersion was
restored immediately and the restored viscosity value was constant and agreed with
the value before applying shear stress.
Moreover, the CNF dispersion also has a cost-benefit advantage compared with SH solution.
The manufacturing cost of the CNF dispersion is much lower than that of SH solution,
because cellulose is abundant in nature. The lower price of the CNF dispersion should
help to reduce the cost of endoscopic treatment. These factors led us to propose that
CNF dispersion could be a better submucosal injection material than 0.4 % SH solution
in terms of its mucosa-elevating capacity, catheter injectability, and the cost of
endoscopic treatment. This is the first preliminary study to demonstrate the potential
usefulness of CNF dispersion for submucosal injection.
Although the present study indicated the superior characteristics of CNF dispersion
over SH solution for submucosal injection, it has some limitations. The major limitation
is that the safety of the CNF dispersion in vivo has not been adequately tested. Our
study was conducted in ex vivo porcine gastric mucosa. The general health, safety,
and environmental aspects of CNF dispersion have recently been evaluated. The results
of toxicity studies suggest that CNF dispersion is not cytotoxic and does not cause
any inflammatory effects in macrophages [10].
In addition, preclinical trials of endoscopic treatment with CNF dispersion in an
in vivo animal model are necessary to confirm the feasibility and safety of CNF dispersion
as a new submucosal injection material. Endoscopic treatments such as EMR or ESD using
TOCN dispersion as a submucosal lifting material will allow us to evaluate the formation
and maintenance of mucosal lesion lifting during endoscopic treatment and the ease
of mucosal resection including cutting speed, hardness, and visibility.
In conclusion, these preliminary experiments have demonstrated the superiority of
CNF dispersion for injection into the submucosal layer and maintenance of mucosal
thickness, illustrating the potential of the CNF dispersion as an ideal submucosal
injection solution. The highly thixotropic CNF dispersion can effectively reduce the
resistance to injection and maintain mucosal elevation for a long period.
