Keywords
traumatic brain injury - aquaporin 4 - neuropeptide Y
Introduction
Traumatic brain injury (TBI) is a common occurrence in road and railway accidents.
Although the exact incidence is unknown, it also is an important occurrence in patients
presenting with domestic injuries like fall from height, domestic violence, and so
forth.[1] The severity of this injury affects the overall survival, prognosis, multiorgan
failure, morbidity, and mortality in such cases. The exact mechanisms that are involved
in this are poorly understood, but inflammation, gastrointestinal (GI) dysfunction,[2] and increased intestinal permeability[3]
[4]
[5] occurring secondary to the TBI have been proposed to play a crucial role in the
pathogenesis. Increased intracranial tension is commonly noted in TBI patients, which
could potentially modulate these changes via the brain–gut axis.[6] GI dysfunction occurs at multiple levels like at the gastric level, duodenal level,
jejunal level, and large intestinal level.[6]
Animal models have shown alterations occurring are mainly in the mucosal structure
causing reduced absorption of nutrients, increased intestinal permeability, translocation
of intestinal bacteria, and endotoxin.[7] The increased permeability leading to faulty transport of solutes and nutrients
across the gut is mainly attributable to the alterations in brain–gut axis mediated
through various molecules like aquaporin 4 (AQ4) and neuropeptide Y (NPY).[8]
[9] Expression of aquaporins, which are water channels expressed in various tissues
mediating entry and exit of water molecules across membranes, has been correlated
with various GI disorders like chronic gastritis, gastric cancer, and so forth.[10] Similarly NPY, which inhibits GI motility and electrolyte secretion in the digestive
tract thus modifying the input to the brain,[8] have also been found raised in TBI induced in animal models.
However, there is lack of clinical studies in humans for the same, as most of these
studies have been performed in animal models under controlled conditions. So there
is limited evidence of TBI-induced changes in human intestines and the pathways involved
in the same.
We intended to understand more about the pathogenesis of TBI induced secondary GI
changes. For this we targeted the jejunum, as it is one of the most important surfaces
in the GI tract responsible for the absorption of nutrients, coming only secondarily
to the duodenum. Another important reason for targeting the jejunum was that there
has been very little research on the jejunal changes in such cases, as the maximum
amount of research has been focused on the gastric and duodenal changes secondary
to TBI. We also intended to find out the expression of AQ4 and NPY in the jejunum
by immunohistochemistry, as markers of TBI-induced GI changes.
This study will help in future to develop the means and medications to limit the GI
damage in such critically ill patients, thus helping them to improve their prognosis
and recuperate to health.
Materials and Methods
-
Study design: It was prospective cohort study in which traumatic head injury patients, older than
16 years, requiring admission to Jai Prakash Narayan Apex Trauma Centre (JPNATC),
All India Institute of Medical Sciences (AIIMS), New Delhi, were considered.
-
Inclusion and exclusion criteria: Cases included patients dying due to traumatic head injury and death occurring in
not less than 24 hours. Controls included patient who had died due to trauma other
than traumatic head injury and also patients who were brought dead to the hospital
after trauma. In patients with blunt/penetrating injury to the abdomen, autopsy was
done 12 hours after the death and any autolysed samples were excluded from cases and
control groups of the study.
-
Sample preparation: Autopsy sections of small intestine 15 cm distal to the ligament of Treitz, at least
10 cm in length, were obtained in 10% buffered formalin solution. Written informed
consent from the authorized relatives was obtained. After receiving in the laboratory,
the samples were grossed for length, color of serosa, presence of perforation, hemorrhage,
fibrosis, scar, exudates, abscesses, sutures, and so forth. Further tissue processing
was done automatically on Histokinette (Leica Biosystems, Germany) automated tissue
processing unit and final embedding in paraffin wax was done on Tissue-Tek (Sakura
Finetek, United States) semiautomated apparatus. Thin sections of 3 to 4 µm were obtained
by sectioning the embedded sample on semiautomated microtome. The sections were put
in a tissue flotation bath and then taken on slides.
-
Staining for light microscopy for jejunum histology: For light microscopy, sections were taken on Superfrost slides (Fisher Scientific,
United States). Then the slides were warmed to 60 to 70°C for 2 minutes on the slide
warming table. This fixed the tissue to the slides. They were stained with hematoxylin
and eosin stain. The light microscopic findings were evaluated by three observers
based on presence or absence of thrombosis, ulceration, and inflammation, and on the
intensity of thinning and/or infarction of wall. Each sample was given a score out
of 7 as per [Table 1] (also refer to [Fig. 1]).
-
Staining for immunohistochemistry for AQ4 and NPY in jejunum: For immunohistochemistry the sections were taken on aminopropyltriethoxysilane (APES)
solution–coated Polysine microslides. These sections were run on BenchMark ULTRA automated
immuno-histo chemistry/in-situ hybridization (IHC/ISH) slide staining system (Ventana
Medical Systems, Inc., F. Hoffmann-La Roche Ltd; automated immunohistochemistry machine)
for staining for AQ4 and NPY. The following are the steps that were performed automatically
in the system. Antigen retrieval was done using cell conditioning 1 buffer in three
steps. First conditioning was done for 8 minutes at room temperature, followed by
mild conditioning for 30 minutes, and finally standard conditioning for 60 minutes.
For AQ4, 1:100 dilutions of AQ4 primary antibodies were applied and incubated overnight
at 4°C followed by washing thrice with 1X PBS (phosphate buffered saline). For NPY,
1:600 dilutions of NPY primary antibodies were applied and incubated at 37°C for 2
hours followed by washing thrice with 1X PBS. Respective horseradish peroxidase-conjugated
secondary antibody for AQ4 and NPY at 1:400 dilutions was applied and washing was
done thrice followed by addition of chromogenic substrate (diaminobenzedine [DAB]).
Slides were dried and then analyzed microscopically.
Table 1
Scoring of microscopic findings
|
Finding
|
Score
|
|
0
|
1
|
2
|
|
Presence of thrombosis
|
Absent
|
Present
|
–
|
|
Presence of ulceration
|
Absent
|
Present
|
–
|
|
Presence of inflammation
|
Absent
|
Present
|
–
|
|
Thinning of wall
|
Nil
|
Mild/moderate
|
Severe
|
|
Infarction of wall
|
Nil
|
Mild/moderate
|
Severe
|
Fig. 1 Histopathological (light microscopic) changes in the jejunal mucosa. (A) Normal morphology of the human jejunal mucosa; H&E stain at 10× power. (B) Presence of inflammation (arrow) seen in the jejunal mucosa; H&E stain at 10×. (C) Infarction (arrow) seen in the jejunal mucosa; H&E stain at 10× power. (D) Thinning of the wall (arrow) seen in the jejunal mucosa; H&E stain at 10× power.
(E) Presence of thrombosis (arrow) seen in the jejunal mucosa; H&E stain at 10× power.
(F) Presence of ulceration (arrow) seen in the jejunal mucosa; H&E stain at 10× power.
(G) Scatter plot showing individual values of cases and controls along with the mean
and standard deviation.
IHC staining for AQ4 and NPY was evaluated blind by three observers and the score
was average for each sample. Slides were examined for the presence or absence of staining,
location of cells showing a positive reaction, and the intensity of positivity of
the reaction in the cells. Accordingly, the slides were graded as 0, 1, 2, 3, and
4, with 0 indicating no reaction and 1, 2, 3, and 4 indicating increasing grades of
positivity.
Results
A total of 40 autopsy samples were obtained for the study which included 20 cases
and 20 controls. One case had to be excluded from the analysis as focused assessment
with sonography for trauma (FAST) was positive. Similarly, three controls also had
to be excluded as one of them was FAST positive, one had a retroperitoneal hematoma,
and one had thoracic and abdominal injury but FAST could not be traced. So, a total
of 19 cases and 17 controls were analyzed in the study. Among the controls, nine were
patients who were brought dead (53% controls approximately). The results are summarized
in the flowchart in [Fig. 2].
Fig. 2 Flowchart of overall case and control selection. FAST, focused abdominal sonography
in trauma; TBI, traumatic brain injury.
Demographics
The median age of the cases was 35 years (range: 19–75 years) and that of the controls
was 39 years (range: 18–75 years). Males outnumbered females in both cases and controls
(16 males against 3 females in cases; 15 males against 2 females in controls). The
leading cause of trauma was road traffic accidents (13/19 in cases and 9/17 in controls).
The other causes were fall from height, railway track accidents, gunshot injury to
head, and so forth. The time from injury to the death of the patient was calculated
in days. The time from injury to death ranged from 1 to 73 days in cases and from
0 to 86 days in controls. The glasgow coma score (GCS) and injury severity score (ISS)
were evaluated at the time of admission to grade the severity of injury. The mean
GCS was decreased in both the cases and the controls, which was statistically insignificant
(cases, 6.36 ± 0.75; controls, 7.52 ± 1.32; p-value = 0.43). The mean ISS was significantly higher in the controls than the cases
(cases, 31.16 ± 2.32; controls, 52.18 ± 5.65; p-value 0.001). The GCS was decreased and the ISS were raised in the cases. Almost
half of the controls were patients who were brought dead, in whom the GCS was lowest
and the ISS was highest. Only two cases had evidence of sepsis. None of the controls
had sepsis.
Gross Findings
The samples were grossed for length, color of serosa, perforation, hemorrhage, fibrosis,
scar, exudates, abscesses, sutures, and so forth. Samples were labeled as “YES,” indicating
significant gross examination finding and “NO,” indicating a normal sample or a sample
with no significant gross findings; 9 cases out of 19 and 8 controls out of 17 had
gross findings. The results obtained are depicted in [Table 2]
[Fig. 3]. The odds ratio was 1.01, which suggests that head injury has a role to play, but
a chi-square test used for analysis indicated that the gross examination findings
were not significantly different between the cases and the controls (p-value > 0.99). This indicates that gross findings can be present post trauma in individuals,
but it cannot reliably differentiate between post TBI and other causes. The individual
details of cases and controls have been enumerated in the [Table 3].
Table 2
Distribution of gross findings in cases and controls
|
Gross findings
|
|
Yes
|
No
|
Total
|
|
Case
|
9
|
10
|
19
|
|
Controls
|
8
|
9
|
17
|
|
Total
|
17
|
19
|
36
|
Table 3
Distribution of gross, microscopic and immunohistochemical expression of AQ4 and NPY
in individual cases and controls
|
Gross findings
|
Microscopic findings score
|
AQ4 intensity
score (0–4)
|
NPY intensity
score (0–4)
|
|
Thrombosis (0–1)
|
Ulceration (0–1)
|
Inflammation (0–1)
|
Thinning (0–2)
|
Infarction (0–2)
|
Total (0–7)
|
|
Cases
|
|
1
|
NO
|
0
|
1
|
1
|
1
|
0
|
3
|
2
|
2
|
|
2
|
YES
|
1
|
1
|
1
|
1
|
0
|
4
|
2
|
4
|
|
3
|
YES
|
0
|
0
|
1
|
0
|
1
|
2
|
2
|
3
|
|
4
|
YES
|
0
|
0
|
0
|
0
|
0
|
0
|
2
|
3
|
|
5
|
YES
|
0
|
0
|
0
|
1
|
2
|
3
|
2.5
|
1
|
|
6
|
NO
|
1
|
0
|
1
|
0
|
0
|
2
|
2
|
4
|
|
7
|
YES
|
0
|
1
|
1
|
0
|
0
|
2
|
2
|
3
|
|
8
|
NO
|
0
|
1
|
1
|
1
|
0
|
3
|
2.5
|
4
|
|
9
|
YES
|
1
|
1
|
0
|
2
|
0
|
4
|
2
|
2.5
|
|
10
|
NO
|
1
|
1
|
0
|
1
|
0
|
3
|
2
|
3
|
|
11
|
NO
|
0
|
0
|
0
|
0
|
0
|
0
|
3
|
2.5
|
|
12
|
NO
|
0
|
0
|
0
|
0
|
0
|
0
|
2
|
1.5
|
|
13
|
NO
|
0
|
1
|
1
|
0
|
0
|
2
|
2
|
1.5
|
|
14
|
YES
|
0
|
1
|
1
|
1
|
0
|
3
|
3
|
2.5
|
|
15
|
NO
|
0
|
0
|
1
|
0
|
0
|
1
|
2.5
|
1.5
|
|
16
|
YES
|
1
|
1
|
1
|
0
|
0
|
3
|
2.5
|
1
|
|
17
|
NO
|
0
|
0
|
1
|
0
|
0
|
1
|
2
|
1
|
|
18
|
NO
|
1
|
0
|
0
|
0
|
0
|
1
|
3
|
1
|
|
19
|
YES
|
1
|
0
|
1
|
0
|
0
|
2
|
3.5
|
1
|
|
Mean
|
Not applicable
|
2.05 ± 0.29
|
2.34 ± 0.10
|
2.26 ± 0.24
|
|
Controls
|
|
1
|
YES
|
1
|
1
|
1
|
1
|
0
|
4
|
3
|
3.5
|
|
2
|
YES
|
1
|
0
|
1
|
1
|
0
|
3
|
0.5
|
2
|
|
3
|
YES
|
0
|
0
|
1
|
0
|
1
|
2
|
3
|
3
|
|
4
|
NO
|
0
|
1
|
1
|
0
|
0
|
2
|
2
|
2.5
|
|
5
|
NO
|
1
|
0
|
0
|
0
|
0
|
1
|
0.5
|
1.5
|
|
6
|
NO
|
1
|
0
|
1
|
0
|
0
|
2
|
3
|
2
|
|
7
|
NO
|
0
|
0
|
0
|
0
|
0
|
0
|
2
|
1
|
|
8
|
NO
|
0
|
0
|
0
|
0
|
0
|
0
|
2
|
1
|
|
9
|
YES
|
1
|
0
|
1
|
2
|
0
|
4
|
1
|
3
|
|
10
|
NO
|
0
|
1
|
1
|
0
|
0
|
2
|
1
|
3
|
|
11
|
NO
|
0
|
0
|
0
|
0
|
0
|
0
|
1
|
3
|
|
12
|
NO
|
1
|
1
|
1
|
0
|
0
|
3
|
1
|
2
|
|
13
|
YES
|
0
|
1
|
0
|
0
|
0
|
1
|
2.5
|
3
|
|
14
|
YES
|
1
|
0
|
1
|
1
|
1
|
4
|
3
|
2.5
|
|
15
|
YES
|
0
|
0
|
0
|
0
|
0
|
0
|
1
|
1.5
|
|
16
|
YES
|
0
|
1
|
1
|
0
|
0
|
2
|
1.5
|
1
|
|
17
|
NO
|
0
|
1
|
0
|
1
|
0
|
2
|
3
|
2.5
|
|
Mean
|
Not applicable
|
1.88 ± 0.34
|
1.82 ± 0.23
|
2.23 ± 0.19
|
|
p-value
|
>0.99
|
0.7
|
0.04
|
0.93
|
Fig. 3 Bar diagram showing the occurrence of gross examination findings among the cases
and the controls.
Microscopic Findings
Changes in the gut mucosa from normal were searched for in the study. For doing this
the standard H&E staining was done. We have graded these microscopic changes on a
score of 0 to 7 based on the presence or absence of thrombosis in the mucosa, or ulceration
and inflammation in the mucosa. Infarction of the mucosal wall and the thinning of
the wall secondarily caused due to it were also noted; 16 out of 19 (~84%) cases and
13 out of 17 (~76%) controls had changes in the jejunal mucosa, and 7 out of 19 cases
and 7 out of 17 controls had the presence of thrombosis. Similarly, 9 cases and 7
controls had ulceration and 12 cases and 10 controls had the presence of inflammation;
7 cases and 5 controls had evidence of mucosal wall thinning, among which 1 case and
1 control had severe wall thinning. Mucosal wall infarction was seen in two cases
and two controls, and one case had severe wall infarction. The mean microscopy score
of cases was 2.05 ± 0.29 and that of controls was 1.8 ± 0.34. An unpaired t-test was used to for comparison. Although microscopic findings were present in both
the cases and the controls, they were not significant statistically (p-value = 0.70). The findings of individual cases and controls are illustrated in [Table 3]
[Fig. 1].
Immunohistochemical Findings
AQ4 and NPY expression in the jejunal mucosa was evaluated by immunohistochemistry
on a score of 0 to 4, with 0 indicating no expression. The mean intensity of AQ4 positivity
in the cases group was 2.34 ± 0.10 and in the control group was 1.82 ± 0.23. An unpaired
t-test was used to compute the difference between the two groups. This difference in
the intensity was statistically significant with a p-value of 0.04, indicating that AQ4 was expressed more in the cases than the controls.
The expression of AQ4 in cases and controls are enumerated in [Table 3]
[Fig. 4]. The mean intensity of NPY expression in the cases group was 2.26 ± 0.24 and in
the control group was 2.23 ± 0.19. An unpaired t-test used to compute the difference between the two groups did not reveal any statistically
significant difference (p-value = 0.93). The expression of NPY in cases and controls are enumerated in [Table 3]
[Fig. 5].
Fig. 4 The categorization of aquaporin 4 by immunohistochemistry of paraffin-embedded human
jejunal tissue with anti–aquaporin 4 antibody (Abcam ab11026) at 1:100 dilution. (A) Grade 0: Nil positivity, at 40× power. (B) Grade 1: Mild positivity, at 40× power. (C) Grade 2: Moderate positivity, at 40× power. (D) Grade 3: High positivity, at 40× power. (E) Grade 4: Intense positivity, at 40× power. (F) Scatter plot showing individual values of cases and controls along with the mean
and standard deviation.
Fig. 5 The categorization of neuropeptide Y by immunohistochemistry of paraffin-embedded
human jejunal tissue with anti–neuropeptide Y antibody (Abcam ab180809) at 1:600 dilution.
(A) Grade 0: Nil positivity, at 40× power. (B) Grade 1: Mild positivity, at 40× power. (C) Grade 2: Moderate positivity, at 40× power. (D) Grade 3: High positivity, at 40× power. (E) Grade 4: Intense positivity, at 40× power. (F) Scatter plot showing individual values of cases and controls along with the mean
and standard deviation.
Discussion
India being a rapidly developing country is prone to vehicular congestion on roads
which is a major factor for RTA, which in turn is a common cause of TBI. As most vehicles
in India are driven by males, they are more likely to be involved in RTA than females.
This same scenario was encountered in our study as RTA was the most common cause of
TBI and the incidence of RTA was more in males than females by a significant fraction.
GI dysfunction is known to complicate the prognosis in patients of TBI, mainly due
to a dysfunction in the brain–gut axis which is formed by the hypothalamus-pituitary-adrenal
axis and the autonomic nervous system. It is responsible for regulating the intestinal
functions, in health as well as in physiologically compromised states.[6]
[11] Dysfunction in this axis results in a myriad of complications like vomiting, gastric
reflux, delayed gastric emptying, intolerance to enteral feeding to reduced intestinal
peristalsis, enterogenous sepsis, and so forth.[12] Various brain–gut peptides like cholecystokinin (CCK), vasoactive intestinal peptide
(VIP), calcitonin gene-related peptide (CGRP), neuropeptide Y (NPY), and substance
P (SP) are known to influence the intestinal mucosa and affect the function and integrity
of remote organs and tissues, leading to systemic inflammatory response syndrome and
multiple organ dysfunction syndromes (MODS).[13]
Studies in animal models under controlled conditions[7]
[14]
[15] have been done to demonstrate the GI changes post TBI, but the same cannot be done
in humans as it would be inhumane to subject humans to such agonizing and life-threatening
trauma. So we designed a study to study the same in patients dying due to TBI. However,
a person could have several pre-existing conditions before the TBI which could also
cause GI changes. To overcome this difficulty a thorough research of the medical history
was undertaken by referring to the medical records, drug history, and by talking to
the relatives and attendants of the deceased. Also patients who underwent any GI intervention
were not included in the study. Utmost care was taken not to include those patients
with significant medical history or conditions that could induce a bias in the study.
As per our understanding, this study design could thus prove useful to understand
the GI changes post TBI in humans.
GCS is a system developed to access the conscious state of a human with the help of
three parameters: eye opening, verbal response, and motor response; the maximum score
being 15, which indicates a fully conscious person and a minimum score of 3 indicates
unconscious and unresponsive person. It has been traditionally used by health care
workers to access and monitor patients with head trauma and patients in intensive
care unit (ICU) settings. The mean score at the time of admission was decreased in
both the groups. A decrease in GCS is expected in the cases group as they are the
ones with TBI. Also, it was the lowest in the patients who were brought dead.
ISS is a scoring system to access the severity of trauma. The score was raised in
the control group because of highest score among the patients who were brought dead.
This score was statistically significant but because of the skewed results no significant
inference could be drawn.
The specimens received for histopathology were inspected under direct vision for morphology
and other significant findings. No significant difference between cases and controls
was found (p-value > 0.99), which indicates some alteration in the morphology of adults can happen
irrespective of TBI. Changes like inflammation, thrombosis, ulceration along with
thinning, and infarction of the intestinal wall that cannot be always discovered by
gross examination need light microscopy.[7] Hence, a score based on this was formulated to grade the severity of the changes.
This score was high in both the cases and controls (p-value = 0.70). Thus, both the groups can be assumed to be comparable as the gross
and microscopic findings were not different between cases and controls. It would be
wise to infer that confounding bias of pre-existing or undetected mucosal pathology
was removed to some extent by these findings. This also means that histopathological
examination could not be used as a standalone test for the assessment of GI changes
in TBI patients, thus implying the necessity of a marker, something which could show
a significant difference in expression among the cases and the controls.
Studies in animals have shown AQ4 and NPY to be raised in jejunal mucosa post TBI.[9] NPY, a 36 amino acid–long polypeptide, is one of the most potent orixegenic peptide
in the body.[16] In the digestive tract, NPY inhibits GI motility and electrolyte secretion, and
in this way modify the input to the brain.[8] On the other hand aquaporins, which are charged integral membrane proteins, regulate
water molecule entry across the cell and block the passage of other ions and solute.[17] They are related to many diseases like diabetes insipidus, salivary gland dysfunction
(Sjogren’s syndrome), cataracts, and even hypertension. A major function of GI tract
of fluid transport is believed to be regulated by AQ4, and so it is accepted that
it plays a role in several GI tract related diseases as well.[18] Thus, AQ4 and NPY have been postulated to play a role in the pathogenesis of intestinal
dysfunction after TBI. Increased NPY levels may be responsible for intestinal ischemia
and hypoxia, whereas AQP4 may play an important role in intestinal edema. The expression
of both AQ4 and NPY has been studied in animals under controlled conditions and found
to be raised and significantly different in TBI.[9] We studied the expression of these two markers, AQ4 and NPY, by immunohistochemistry.
Our study revealed a significantly raised expression of AQ4 in jejunum post TBI (p-value = 0.04; [Table 3]). However, the same was not true for the expression of NPY (p-value = 0.93; [Table 3]). The exact reason for this difference between the findings between animal studies
and this study could not be ascertained.
Conclusion
TBI has a deep impact on the GI morphology and functioning, and ultimately on the
prognosis. Traditional methods like histopathological examination cannot be used to
distinguish between the changes occurring in the jejunal mucosa post TBI and post
non-TBI. Expression of AQ4 in the jejunum post TBI is increased, as demonstrated by
immunohistochemistry, and it can be used as a marker of GI injury of TBI in humans.
The value of NPY expression, although increased post TBI in controlled studies in
animals, needs further evaluation in humans. This study is novel by being the one
of the few to explore the secondary changes in human jejunum induced by TBI. Further
studies will be required to correlate the intensity of expression with the intensity
of TBI.
Limitations of the study: As stated time and again, isolated TBI in humans, wherein
the most common cause of TBI being road traffic accidents, is an idealistic condition
in real world. And unfortunately, it cannot be artificially induced in the laboratory.
So, the possibility of an unknown and undiagnosed injury always lurked around in the
background.
The controls we used were diseased controls as they had some form of trauma or the
other except TBI and abdominal injury. So, comparison with healthy age-matched controls
could not be done.
