Identification of Pattern Recognition Receptors and Quantification of IL-8

Production in HT-29 human intestinal epithelial cells challenged with Vibrio

cholerea and Vibrio parahaemolyticus

Steven Liu & Kelley Zimmerman

Dr. Candace Winstead

Bio 426, Spring 2013

California Polytechnic State University, San Luis Obispo

Introduction

Vibrio species are found naturally worldwide in estuarine and marine environments and

account for a substantial amount of foodborne and waterborne infections particularly in locations

with poor sanitation.  While each Vibrio species uses different mechanisms to cause disease but

all infections resulting from this genus of bacteria lead ultimately to diarrhea. The two vibrio

species used to stimulate the HT29 cells in our study were V. cholerae and V. parahaemolyticus.

V. cholera is the causative agent of cholera which is endemic in places with poor sanitation and

limited access to clean water. There are 140 different serotypes of V. cholerae of which serotype

O1 is often liked to major pandemics. Levels of V. cholerae are observed to be at much higher

levels during warm months and the number of cases correlates during warmer months. There are

a variety of virulence factor associated with V. cholerae such as cholera toxin, toxin co-regulated

pilus, chemotaxis protein, accessory cholera enterotoxin, zonula occludens toxin, and

neuraminidase. The associated biological effect of cholera toxin is the hypersecretion of water

and electrolytes, toxin co-regulated pilus is the binding site for the bacteriophage CTXФ (which

encodes for the cholera toxin) which allows for adherence to intestinal mucosal cells.  The

chemotaxis protein serves as an adhesion factor and the accessory cholera enterotoxin works to

increase intestinal fluid secretions.  The Zonula occludens toxin increases intestinal permeability

while neuraminidase allows for modification of the cell surface to increase GM1 binding sites for

cholera toxin.  It is the massive movement of ions and water from the cells into the gut that

causes the characteristic rice water stools observed with cholera. Infection with V. cholerae can

range from asymptomatic infection to rapidly fatal diarrhea. V. cholerae is non-invasive and

mediates a non-inflammatory infection. There is evidence though that cholera toxin may induce a

TH2 response that causes activation of the humoral immune response and some innate cells (2).

Unlike pathogens that cause invasive diarrhea through the invasive destruction of epithelial cells,

the histopathological features of the intestinal tract are not disturbed during a V. cholerae

infection (2).   While V. cholerae infections are primarily derived from contaminated water V.

parahaemoltyicus is often associated with the consumption of raw shellfish. In addition v.

cholerae tends to have a longer incubation period (2-3 days) compared with V. parahaemolyticus

(mean 24 hours, 5-72 hours) and the mortality rate of untreated patients with symptomatic

cholera is 60% while most infections with V. parahaemolyticus tend to clear up rather easily.  V.

parahaemolyticus is an invasive pathogen that causes a fluid rich diarrhea through disruption of

of the intestinal epithelium which causes substantial inflammation (3). There are also a few

different virulence factors associated with this pathogen including a type III secretion system and

the Kanagawa hemolysin (heat stable, direct hemolysin).  V. parahaemolyticus is also able to

cause an efflux of ions through the use of its hemolysin which causes the secretion of chloride

ions resulting in watery diarrhea.  

In this experiment V. cholerae and V. parahaemolyticus were used to stimulate human

epithelial cells (HT29). HT29 cells produce secretory immunoglobulin A (IgA) and a

carcinoembryonic antigen.  A colon adenocarcinoma was used to establish this HT29 cell line.

Studies have shown that HT29 cells express TLR5 and effectively react to flagellin using the

TLR5 receptor (4).  Other studies have also shown that these cells also express the intracellular

protein NOD1. NOD1 functions in intestinal epithelium as a mechanism to signal activation of

the innate immune system independently of TLR5 during an invasive infection (5).  During

infection HT29 cells have demonstrated an up-regulation in IL-8 production in response to the

cytokine IL-1β and the proinflammatory cytokine TNF-α (6).  We expect to see an upregulation

in IL-8 production in HT29 cells after they are stimulated with Vibrio species due to the flagella

on these species which induce proinflammatory responses and IL-8 production (4).  We also

expect to observe a higher level of IL-8 production in cells stimulated with V. parahaemolyticus

because it causes inflammatory diarrhea in comparison to those challenged with V. cholerae

which results in very minimal inflammatory. We also expect higher NOD1 expression in HT29

cells challenged with V. parahaemolyticus in comparison with cells stimulated with V. cholerae.

We expect this result based on results from a study that demonstrated higher levels of TNF-α in

plasma and stool samples from those infected with V. parahaemolyticus than those infected with

V. cholerae which were at undetectable levels (7).

Materials and Methods

Immunfluorescence phenotyping of HT-29 intestinal epithelial cells. HT-29 human

intestinal epithelial cells were adhered to coverslips and fixed depending on the location of the

pathogen recognition receptor. A moist chamber was set up for each stain using a sectored petri

dish with wet Kimwipes in each sector. The coverslip with the adhered cells were placed on a

piece of parafilm cell side facing up. TLR-5, a cell surface receptor, was fixed with 50 µL of a

3.7% formaldehyde solution to stabilize surface proteins and allowed to incubate for 10 minutes.

NOD-1, an intracellular receptor, was fixed with a 3.7% formaldehyde/0.1% TritonX solution

and allowed to incubate for 10 minutes. TritonX was necessary for to permeabilize the

membrane so the antibody could enter the cell. The formaldehyde solution was aspirated and the

coverslips were washed 3X with 200 µL of PBS and allowed to sit for one minute before

repeating. A 1:50 dilution each of anti-human (positive control), non-binding goat IgG (negative

control), anti-human TLR-5/goat and anti-human NOD-1/goat primary antibody in PBS with 1%

blocking serum was added to a clean parafilm sector. The coverslip containing the HT-29

epithelial cells were flipped cell side down onto the stain and allowed to incubate for 45 minutes

at room temperature. At the end of the incubation period, the primary antibody was aspirated and

the coverslip was washed 4X with 1000 µL of PBS and allowed to sit for one minute before

repeating. Following the primary stain, 50 µL of the fluorochrome conjugated secondary

antibody, anti-goat IgG- Texas Red/rabbit, at a 1:50 dilution was added to a clean parafilm sector

and the HT-29 epithelial cells were flipped cell side down onto the stain and allowed to incubate

in the dark for 45 minutes. At the end of the incubation period, the secondary antibody was

aspirated and the coverslip was washed 4X with 1000 µL of PBS and allowed to sit for one

minute before repeating. A slide containing 5 µL of anti-quench reagent (AQ with DAPI) was

prepared and the coverslips were inverted cell side down onto the reagent. The slides were stored

at -4°C and examined under a fluorescence microscope with the appropriate filters approximately

six days later.

HT-29 cell culture and stimulation. HT-29 intestinal epithelial cells were grown to

100% confluency (1.04 x 106 cells/flask) and a T-75 flask was seeded at a similar density as the

HT-29 intestinal epithelial cells. The cultures were started overnight in 3 mL of TSB with V.

cholera and V. parahaemolyticus using a 1 µL calibrated inoculating loop. The bacterial cells

were washed once with PBS and then centrifuged at 3,000 RPM and then followed by

resuspension in fresh MEM with no serum. The washed HT-29 cells were washed with PBS and

challenged with a MOI of ~100 Vibrio sp. per HT-29 cell for two hours. Supernatant was

harvested by microfuging the cell cultures at 3,000 RPM to remove extraneous material.

Protein extraction preparation. Adherent and non-adherent cells were washed with ice-

cold PBS. The non-adherent cells were centrifuged at 800-1000 RPM for 5 minutes to collect the

cells. Ice-cold modified RIPA buffer (plus protease inhibitors) was added to the cells and

adherent cells were scraped off with a cooled scraper and transferred into a centrifuge tube. The

suspension was placed in an orbital shaker in the cold room for 15 minutes to facilitate lysis of

the cells. The resulting lysate was centrifuged at 14,000 xg in a precooled centrifuge for 15

minutes and then the supernatant was transferred to a fresh centrifuge tube and stored at -20°C.

Quantitation of protein extracts from stimulated HT-29 intestinal epithelial cells.

The quantitation of proteins from supernatants of cell lysates after challenge with V. cholerae

and V. parahaemolyticus was collectively performed by all the students using the BCA assay.

One group of students prepared the BSA standards, one group prepared the working reagent and

another group loaded the unknown proteins onto the microplate. Our group was responsible for

the preparation of diluted albumin (BSA) standards. Nine vials (A-I) were prepared and the first

three vials (A-C) were prepared by adding 300, 375 and 325 of µL source BSA. The next five

vials (D-H) were prepared with a dilution of 175 µL vial B dilution, 325 µL vial C dilution, 325

µL vial E dilution, 325 µL vial F dilution and 100 µL vial G dilution. The last vial (I) did not

contain any source BSA. The BCA working reagent and diluted albumin standards were mixed

and then loaded onto a micoplate. The standards were loaded in triplicate and the protein lysate

was added singly to each well of the microplate. The absorbance of the proteins was measured at

562 nm by plate reader. Average absorbance measurement of the blank standard replicate was

subtracted from the absorbance values of all individual standard and lysate sample replicates. A

standard curve was generated by plotting the average blank-corrected absorbance of each BSA

standard against its concentration (µg/mL) and used to determine the protein lysate

concentrations.

Western immunoblot. Protein lysates and 5X sample buffer were added together and

placed in a heating block for 10 minutes and then immediately placed on ice. A gel apparatus

was set up for a 10% polyacrylamide gel and an appropriate amount of 1X SDS-PAGE running

buffer was added. The wells of the gel were loaded with 5 µL of molecular weight marker and

the calculated volumes of each sample were loaded into each well using special gel loading tips

that were able to be inserted between the two glass plates. Samples were loaded in duplicate to

obtain a gel with two identical halves. The SDS-PAGE gel was run for approximately 60-80

minutes at a voltage of 120 V until the dye reached the bottom of the gel. The gel was then

sandwiched between stacks of midi membranes in a cassette. The anode stacked was placed

membrane side up first at the cassette base and the gel was aligned on the membrane and then the

cathode stack was placed on top of the gel. A glass pipette was rolled over the anode and cathode

membranes to remove any air pockets. The cassette was placed in the Trans-Blot Turbo

instrument to complete the transfer. Resulting PVDF membranes were rinsed twice with

deionized water, cut in half and stored in a 15 cc tube filled with PBS on a rotator. Following

blocking by the instructor with 5% nonfat milk in PBST, the membranes were washed 3X with

PBS+0.05% TWEEN rotating for 5 minutes each time. One of the membranes were incubated

with 3 mL of a 1:500 rabbit α human NOD-1 in 20 mL 1:10 BLOTTO and the other membrane

was incubated with 3 mL of a 1:500 rabbit α human TLR-5 in 20 mL 1:10 BLOTTO. The

membranes were allowed to incubate for 48 hours at 4°C on a rotator. After incubation, the

membranes were washed again in PBS+0.05% Tween and incubated for an hour with a

secondary antibody of 1:5000 goat α rabbit. A mix detection reagent of horse radish peroxidase

conjugated probes in 20mL 1:10 BLOTTO was mixed with each of the substrates. Blots were

incubated for 1 minute with the mix detection reagent and excess reagent was drained. The blots

were then wrapped in plastic wrap protein side down and imaged with a BioRad fluorimager to

detect luminescence.

Vibrio sp. challenge of HT-29 cells and quantitation of response by IL-8 ELISA. 100

µL of mouse α human IL-8 diluted in coating buffer (0.1 M sodium carbonate, pH 9.5) was

added to each well of the ELISA plate, sealed and allowed to incubate at 4°C. The instructor

aspirated and washed the wells 3X with 300 µL/well wash buffer and then blocked with 200

µL/well assay diluent and allowed to incubate at room temperature for 1 hour. The wells were

aspirated and washed again with Wash Buffer (1X PBS+0.05% Tween 20). The standard and

sample dilutions were prepared in Assay Diluent (1X PBS+10% FBS). 100 µL of standard,

sample and control were added to the appropriate wells. The standards were performed in

triplicate and diluted across the plate by starting with 200 µL of the 500 pg/mL standard and then

removing 100 µL of the standard and transferring it to the next well with 100 µL of assay

diluent. Standards were diluted in this manner until the final dilution contained 0.25 pg/mL. All

other samples were performed in duplicate and the plate was sealed and incubated for 24 hours at

room temperature. A working detector reagent was provided which contained a dilute detection

antibody (biotinylated anti-human IL-8) at a 1:250 dilution in assay diluent and a dilute enzyme

reagent (streptavidin-HRP) at 1:250 dilution. 100 µL of the working detector was added to each

well and the plate was sealed and incubated for 1 hour at room temperature. The wells were

aspirated and washed 7X with 300 µL of wash buffer. 100 µL of substrate solution was added to

each well and incubated in the dark at room temperature for 15 minutes. 50 µL of stop solution

was added to each well and then the absorbance was measured at 450 nm. A standard curve of

the absorbance vs. IL-8 concentration was generated from the mean absorbance data of the

standards to determine IL-8 concentration.

Results

HT-29 human intestinal epithelial cells express NOD-1 and TLR-5.

Immunofluorescent phenotyping of HT-29 human intestinal epithelial cells were selected for the

cell surface pattern recognition receptors TLR-5 and NOD-1. TLR-5 and NOD-1 primary

antibody staining and DAPI nuclear staining was viewed with fluorescence microscopy. Goat

anti-human TLR-5 stain (Fig. 1b) exhibited very small amounts of fluorescence to none at all of

intestinal epithelial cells. Similarly, the negative control of non-binding goat IgG (Fig. 1d) also

exhibited very small amounts of fluorescence to none at all of intestinal epithelial cells. The

DAPI staining for TLR-5 and the negative control of TLR-5 (Fig. 1a/c) exhibited some

fluorescence of intestinal epithelial cells, but the fluorescence emitted is relatively weak.

(a) (b) (c) (d)

Figure 1. Immunofluorescent and nuclear staining of HT-29 cells for presence of TLR-5. (a) represents an AQ DAPI stain and (b) represents a goat α human TLR-5 stain. (c) represents an AQ DAPI stain and (d) represents a non-binding goat IgG (negative control) of TLR-5.

(a) (b) (c) (d)

Figure 2. Immunofluorescent and nuclear staining of HT-29 cells for presence of NOD-1. (a) represent an AQ DAPI stain and (b) represents a goat α human NOD-1 stain. (c) represents an AQ DAPI stain and (d) represents a non-binding goat IgG (negative control) of NOD-1.

Goat anti-human NOD-1 stain exhibited a bright fluorescence (Fig. 2b) of intestinal epithelial

cells. Similary, the negative control of non-binding goat IgG for NOD-1 (Fig. 2d) also exhibited

a small but bright fluorescence of a single intestinal epithelial cell. The DAPI staining for NOD-

1 and the negative control for NOD-1 exhibited some fluorescence. DAPI staining of NOD-1

(Fig. 2a) reveals a semi-bright cell but DAPI staining of the negative control of NOD-1 (Fig. 2c)

reveals a much dimmer fluorescence.

HT-29 cells exhibit greater IL-8 production when challenged with V.

parahaemolyticus. HT-29 human intestinal epithelial cells were challenged with V. cholerae and

V. parahaemolyticus in order to quantify IL-8 concentrations. The IL-8 concentrations were

determined by performing an ELISA and constructing an IL-8 standard curve as shown in Fig. 3.

0 50 100 150 200 250 300 350 400 450 5000

0.20.40.60.8

11.21.41.61.8

2

f(x) = 0.00326526025328486 x + 0.15751324586703

IL-8 Standard Curve

IL-8 Concentration (pg/mL)

Ab-sorbance450nm

Figure 3. Plot of IL-8 standard curve generated from IL-8 concentration and measured

absorbance values at 450 nm.

Higher IL-8 concentrations are positively correlated with higher absorbance values. Data points

that expressed higher IL-8 concentration and absorbance values were subject to greater deviation

from the regression line, while data points that expressed lower IL-8 concentrations and

absorbance values were very close to the line of best fit. The IL-8 concentrations of the unknown

samples (Table 1) were calculated from the linear equation of the IL-8 standard curve (y =

0.0033x + 0.1575 where y represents absorbance and x represents IL-8 concentrations).

Table 1. IL-8 Concentrations of V. parahaemolyticus, V. cholerae and negative control

Sample Absorbance450nm Concentration (pg/mL)Vp1 1.64 435.24Vp2 1.60 422.59Vp3 1.75 468.47Vc1 1.10 277.00Vc2 1.11 279.79Vc3 1.10 374.79(-)1 0.197 10.530(-)2 0.197 10.800(-)3 0.190 8.320

V. parahaemolyticus produced a mean IL-8 concentration of 442.10 pg/mL and V. cholerae

produced a mean IL-8 concentration of 277.20 pg/mL and the negative control produced a mean

IL-8 concentration of 9.74 pg/mL. V. parahaemolyticus invoked significantly greater production

of IL-8 than V. cholerae in response to challenge by the Vibrio sp as shown in Figure 4. Both

Vibrio sp. produced significantly greater concentrations of IL-8 than the negative control.

V. parahaemolyticus V. cholerae (-) control0

50

100

150

200

250

300

350

400

450

500

IL-8 Quantitation by ELISA

Sample

Concentration (pg/mL)

Figure 4. Column chart of concentration of IL-8 (pg/mL) production when challenged by V. parahaemolyticus, V. cholerae and negative control with standard error bars.

Western immunoblot. The BCA standard curve was used to determine the protein

concentrations of the various cell lysates. The standard curve (Fig. 5) shows very little variation

of the data points from the regression line.

Figure 5. A BCA protein assay standard curve generated from known concentrations of BSA. The equation of the line was used to determine the protein concentrations of the various protein lysates used in the Western blot.

β-actin and NOD-1 probed Western blots failed to produce any results. As Fig. 6 indicates, no

bands were visible on the blots and only a grainy image was obtained.

Figure 6. Western blot analysis of NOD-1 and β-actin. It is impossible to determine the location of the lanes and visibility of any bands is minimal due to the grainy quality of the image obtained.

Western blot of TNF-α was provided by the instructor (Fig. 7). GAPDH was not stimulated by

either V. parahaemolyticus or V. cholerae because it served as the negative control. GADPH

bands can be observed in all lanes at approximately 45 kD. The TNF-α band, approximately 17

kD, were only visible in lanes 7-9 which corresponded to V. cholerae stimulated HT-29

epithelial cells. Both GAPDH and TNF-α were probed on the same gel.

Figure 7. Western blot analysis of GAPDH and TNF-α. Lanes 1-3 contain the negative control GAPDH (negative control), lanes 4-6 contain V. parahaemolyticus stimulated cells, lanes 7-9 contain V. cholerae stimulated cells and lane 10 contains a molecular weight marker. GAPDH and TNF-α were probed on the same gel.

Discussion

Immunofluorescent staining demonstrated that HT-29 human intestinal epithelial cells

express the pattern recognition receptors TLR-5 and NOD-1. TLR-5 is present on the surface of

HT-29 epithelial cells as evidenced by the DAPI stain and fluorescence. NOD-1 was also present

in the cytosol as shown by the DAPI stain and fluorescence. Collectively, this indicates that the

primary antibodies were able to bind to both TLR-5 and NOD-1. However, it should be noted

that observation with fluorescence microscopy was performed six days after addition of the

fluorochrome conjugated secondary antibodies and DAPI stain which may have affected the

results. The fluorescence detected from both TLR-5 and NOD-1 epithelial cells were only faintly

fluorescent which may suggest fading of the fluorochromes and DAPI stain. This error was

observed in all of the stains performed. The negative controls were not helpful in distinguishing

whether the HT-29 epithelial cells expressed TLR-5 or NOD-1 because they also showed signs

of staining and fluorescence. This is most likely due to an error introduced in the staining

procedure. The negative control may have been saturated with fluorochrome conjugated

secondary antibodies and DAPI such that it also emitted fluorescence. Therefore, we cannot

conclude with confidence that the HT-29 epithelial cells preferentially express one pattern

recognition receptor over another because they all appear alike.