Serum Amyloid A-Luciferase Transgenic Mice: Response to Sepsis, Acute Arthritis, and Contact Hypersensitivity and the Effects of Proteasome Inhibition

Construction of pSaa1-luc vector and generation of Saa1-luc transgenic mice

A mouse SAA1 bacterial artificial chromosome clone was obtained through PCR screening of a mouse C57BL/6J genomic DNA library (Research Genetics). A promoter fragment of 7.7 kb was cloned into the pGL3Basic vector (Promega) upstream of the firefly luciferase gene sequence, and the construct was designated pSaa1-luc. The Saa1-luc transgene cassette was released from the vector through double digestions with NotI and SalI and was used to generate transgenic mice in the BALB/C background strain with standard pronuclear microinjection techniques. The transgenic model described in this paper is named BALB/C-Tg (Saa1-luc)Xen and is abbreviated Saa1-luc throughout the text.

Reagents

We purchased bacterial LPS (from Salmonella abortus equi) and zymosan A (a cell wall preparation from Saccharomyces cerevisiae) from Sigma-Aldrich, rTNF-α and IL-1β from R&D Systems, and bortezomib (velcade, PS-341) from Millennium Pharmaceuticals.

In vivo imaging of luciferase activity

In vivo imaging was performed using an IVIS imaging system (Xenogen) as previously described (25). Saa1-luc transgenic mice were anesthetized with isoflurane and injected i.p. with 150 mg/kg luciferin (Biosynth). Ten minutes after the luciferin injection, mice were imaged for 1–60 s. Photons emitted from specific regions were quantified using LivingImage software (Xenogen). In vivo luciferase activity is expressed as photons per second.

Induction of SAA1 promoter-driven luciferase expression in transgenic animals

Induction of SAA1 promoter-driven luciferase expression was followed after i.p. injection of LPS (1 mg/kg), IL-1β (1 μg/mouse), or TNF-α (2 μg/mouse) into Saa1-luc transgenic mice, 3–6 mo of age. Control mice were injected with saline. At selected time points after the treatment, mice were i.p. injected with luciferin and were imaged 10 min later with the IVIS imaging system described above. To test the effect of bortezomib on LPS-triggered luciferase expression, mice were pretreated with bortezomib (1 mg/kg i.v.) 1 h before LPS injection, and the luciferase signal was monitored through imaging.

Tissue luciferase activity

Selected mouse organs were removed, homogenized in 3 vol of PBS containing a protease inhibitor mixture (Roche), and lysed with passive lysis buffer (Promega). After centrifugation at 14,000 rpm for 10 min at 4°C, the supernatant was collected. Luciferase activity was assayed using the Luciferase Assay System (Promega) and a TD 20/20 luminometer (Turner Design). Protein concentration was estimated with Bradford reagent (Sigma-Aldrich).

Acute arthritis model by intra-articular administration of zymosan

Zymosan A was suspended in sterile saline containing 5% glucose at a concentration of 30 mg/ml. Saa1-luc mice, 3–6 mo of age, were anesthetized with isoflurane, the hind legs of the mice were shaved, and the skin was sterilized with 70% ethanol. The right knee tendon was exposed, and 10 μl of zymosan A suspension was injected intra-articularly through the tendon with a 25-gauge needle. The left rear knees of the same mice were intra-articularly injected with 5% glucose in saline (sham control). Mice were imaged at selected time points after the injection. To test the effect of bortezomib on zymosan-triggered luciferase expression, mice were pretreated with bortezomib (1 mg/kg i.v.) 1 h before zymosan administration, and the luciferase signal was monitored through imaging.

Induction of contact hypersensitivity (CHS) reactions

Male Saa1-luc mice, 3–6 mo of age, were shaved in the thorax area and treated topically with 50 μl of 2% oxazolone solution (in acetone/olive oil (4/1, v/v)). Control mice were treated with vehicle (acetone/olive oil (4/1)) alone. Bioluminescent images were collected at various time points after treatment.

Measurement of SAA1 in serum and tissues by ELISA

A mouse SAA1 immunoassay kit (BioSource International) was used to measure SAA1 level in the serum. Both male and female Saa1-luc mice (n = 6), 3–6 mo of age, were injected with LPS (1 mg/kg i.p.) or TNF-α (2 μg/mouse i.p). Blood samples were collected 4 h after the treatment. Before the ELISA, serum samples from LPS- or TNF-α-treated mice were diluted 1/1000 in sample diluent. Control serum samples were diluted 1/10 in sample diluent. Samples were measured for the SAA1 concentration following the manufacturer’s protocol.

We also measured local SAA1 production in the zymosan-injected knee joint or the oxazolone-treated skin area. The knee joint or skin tissue was harvested and ground into fine power while frozen with liquid nitrogen, followed by brief sonication in 3 vol of PBS containing a protease inhibitor mixture (Roche). The homogenate was additionally lysed with passive lysis buffer (100 μl of homogenate, 300 μl of PBS, and 100 μl of 5× passive lysis buffer (Promega)). After centrifugation at 14,000 rpm for 10 min at 4°C, the supernatant was collected for measurement of SAA1 concentration, luciferase activity, and total protein, as described above. No dilution was needed for the ELISA.

Northern blot analysis

Total RNA was isolated from mouse tissue using RNAwiz (Ambion) and further purified using an RNeasy kit (Qiagen). A total of 5 μg of RNA was analyzed by Northern blot using a NorthernMax system (Ambion). A 25-nt SAA1-specific probe (5′-CCTTGGAAAGCCTCGTGAACAAATG-3′) was used for hybridization to the RNA filter at 37°C in ULTRAhyb-Oligo hybridization buffer (Ambion). After hybridization, the signal was detected using a BrightStar BioDetect kit (Ambion).

Statistics

Nonparametric tests for significance were used to test whether changes in luciferase signal from baseline were significantly greater than zero (sign test) and whether the changes from baseline were significantly different between treatment groups (Mann-Whitney U test). Values are presented as the mean ± 1 SE in the graphs and text unless otherwise noted. For some statistical tests, genders were combined to increase sample number in each group. All significance levels are two-sided.

Induction of SAA1 promoter-driven luciferase expression by LPS and TNF-α in Saa1-luc transgenic mice

We generated Saa1-luc transgenic founders whose progeny were screened for luciferase expression in response to LPS as described in Materials and Methods. All founder lines showed robust luciferase induction in the liver after LPS treatment. One founding line was selected for the studies described in this report. Fig. 1⇓ shows the results of treatment with LPS, TNF-α, and IL-1β. Luciferase expression was detected by bioluminescent imaging. LPS treatment caused a robust induction of luciferase signal in the hepatic region of the transgenic mice. Induction was detectable 2 h after LPS treatment and reached a peak at 4–7 h. The signal was reduced sharply at 24 h, and by 48 h the signal had returned to the baseline level. Control mice that were injected with saline did not show any induction (results not shown).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

In vivo IVIS imaging (Xenogen) of anesthetized 3- to 6-mo-old Saa1-luc transgenic male and female mice after i.p. injection of 150 mg/kg luciferin. The mice were treated with LPS (1 mg/kg; top panel), TNF-α (2 μg/mouse; middle panel), or IL-1β (1 μg/mouse; bottom panel) and imaged at 0, 2, 4, 7, 24, and 48 h after treatment. Representative mice from each treatment group are shown. The color overlay on the image represents the photons per second emitted from the animal in accord with the pseudocolor scale shown next to the images. Red represents the highest number of photons per second, whereas blue represents the lowest number of photons per second. After all three treatments, luciferase activity was clearly induced in the region of the liver.

Luciferase signal from the liver region was quantified using LivingImage software to produce the data shown in Fig. 2⇓ (upper panel). Male mice had significantly higher mean luciferase signals than female mice at baseline (mean ± SE, 142.7 ± 62.6 vs 25.8 ± 24.9 million photons/s, respectively; p < 0.0001). After LPS treatment, both males and females showed significant induction of luciferase signal at all time points compared with the baseline level. In relation to the baseline, the fold increase in luciferase signal was similar in males and females (Fig. 2⇓, right top panel). Male mice showed a peak of luciferase induction by 4 h, whereas female mice had a peak of induction at 7 h. At the peak, the luciferase signal was induced ∼2400-fold in both sexes (difference between sexes, p = 0.57). Neither sex exhibited induction after saline injection (results not shown). In another experiment, we found that there was no difference by age for luciferase induction after LPS injection among mice aged 2–7 mo (data not shown).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

Change in luciferase activity over time, as expressed by the difference between baseline photons per second and subsequent measurements (left panel), and as multiples of the baseline activity (right panel). Values are the mean increase ± SE or the fold increase ± SE. Nonparametric significance levels from the sign test: ∗, 0.01 < p ≤ 0.05; §, p < 0.0001. In the top panels, 18 male and 14 female Saa1-luc transgenic mice were measured at baseline and 4, 7, and 24 h after LPS injection. Two subsets of mice treated with LPS received supplementary imaging at 2 h (seven males and three females) and 48 h (six males and six females) after LPS treatment. In the middle and lower panels, Saa1-luc transgenic mice (three males and three females) were treated with TNF-α (2 μg/mouse i.p) or IL-1β (1 μg/mouse i.p.) and imaged at 2 h (TNF-α only) and 4, 7, 24, and 48 h after treatment. When the sexes were combined for statistical analysis, there were significant fold increases relative to baseline at all time points after both TNF-α and IL-1β treatments (p = 0.03).

Robust luciferase induction was also observed in the hepatic region of both sexes after i.p. injection of either TNF-α (Fig. 1⇑, middle panel) or IL-1β (Fig. 1⇑, bottom panel). The middle and lower panels of Fig. 2⇑ display the corresponding kinetics of luciferase induction. The luciferase signal was readily detectable 2 h after TNF-α injection, reached a peak at 4 h, started to decline at 7 h, and by 48 h was only marginally above the baseline. The fold increases in luciferase signal for the sexes combined (n = 6/group) were significantly greater than zero at all time points (p = 0.03; Fig. 2⇑). The kinetics were similar for hepatic luciferase induction by IL-1β. The fold increase in luciferase signal after treatment with IL-1β showed no statistically significant difference between sexes. When data from both sexes were combined, there was a significant positive increase induced by IL-1β 4, 7, 24, and 48 h after treatment (p = 0.03; Fig. 2⇑).

Induction of SAA1 promoter-driven luciferase expression by LPS and TNF-α corresponded with the induction of serum SAA1

We analyzed the serum level of SAA1 after LPS or TNF-α treatment. Mice were bled 4 h after treatment, and the SAA1 serum concentration was tested by ELISA. The concentrations of SAA1 in untreated control mice (mean ± SE) were 0.6 ± 0.1 and 0.7 ± 0.1 μg/ml in males and females, respectively. The SAA1 level was raised to an average of 1793 ± 117 and 1318 ± 315 μg/ml in male LPS- and TNF-α-treated mice, respectively, representing 2846- and 2092-fold inductions. In female mice, the SAA1 levels were raised to averages of 1435 ± 148 and 1232 ± 249 μg/ml after LPS and TNF-α treatment, respectively, representing 2119- and 1820-fold inductions for LPS and TNF-α, respectively. The mean serum SAA1 levels in LPS- and TNF-α-treated mice were all significantly greater than the control values (p = 0.004), and the fold increases in serum SAA1 induction measured by ELISA agreed well with those obtained by bioluminescent imaging.

SAA1 promoter-driven luciferase expression is induced in multiple tissues after LPS and TNF-α treatment

We analyzed luciferase activity in selected organs in Saa1-luc mice 4 h after i.p. injection of LPS, TNF-α, or saline. Ex vivo luciferase activity was detected in the liver and in all other dissected organs of LPS-, TNF-α-, and saline-injected mice (Table I⇓). However, induction of luciferase activity was significantly greater in the LPS- and TNF-α-treated than in the saline-injected mice for liver, lungs, kidneys, brain, and heart. In the intestine, LPS produced significant induction of luciferase in both male and female tissues, whereas TNF-α produced significant induction in male, but not female, intestinal samples. In the spleen, induction was significant after both LPS and TNF-α treatments in male, but not female, mice. As calculated from the mean of the saline-injected mice, LPS treatment in male and female mice caused luciferase induction of 4508- and 5762-fold, respectively, in the liver, 489- and 208-fold in the lungs, 68- and 30-fold in the small intestines, 71- and 75-fold in the kidneys, 1.6- and 1.5-fold in the brain, 68- and 21-fold in the heart, and 2.1- and 1.6-fold in the spleen. The corresponding values for TNF-α treatment in male and female mice were inductions of 3016- and 2288-fold, respectively, in the liver, 67- and 18-fold in the lungs, 329- and 4-fold in the small intestines, 19- and 56-fold in the kidneys, 0.7- and 0.7-fold in the brain, 53- and 12-fold in the heart, and 1.5- and 1.7-fold in the spleen.

Consistent with the in vivo results at 4 h in Fig. 1⇑, the ex vivo analyses described above of luciferase induction in selected organs showed that males had significantly greater induction than females in the liver, lungs, intestines, and heart (p = 0.05) after either LPS or TNF-α treatment. Despite differences in the amount of induction, male and female LPS-treated mice showed a similar pattern of increased luciferase expression across organs. Liver demonstrated by far the greatest increase in luciferase expression in both sexes, whereas brain and spleen demonstrated the smallest increases. In male and female LPS-treated and TNF-α-treated mice, the luciferase activity in all other tissues was significantly less than that in the liver (Table I⇑).

Induction of SAA1 promoter-driven luciferase expression by LPS is inhibited by bortezomib

The presence of NF-κB-responsive elements in the SAA1 promoter suggests that the NF-κB signaling pathway is involved in the regulation of SAA1 expression. We studied the effect of bortezomib on SAA1 promoter-driven luciferase induction by LPS. Bortezomib is a proteasome inhibitor that inhibits IκB degradation and therefore suppresses NF-κB-mediated induction of transcription. Saa1-luc mice were pretreated with bortezomib (1 mg/kg i.v.) 1 h before LPS treatment. Fig. 3⇓ (top panel) shows that compared with Saa1-luc mice treated with LPS alone, the bortezomib-cotreated Saa1-luc mice showed less induction of luciferase signal in the hepatic region. Fig. 4⇓A shows that luciferase induction was observed 2 h after LPS treatment and reached a peak at 4 h. At the peak, LPS-treated mice showed a 2888 ± 764-fold induction of hepatic luciferase signal. Mice pretreated with bortezomib showed only a 301 ± 50-fold induction at 4 h, which represented a 90% inhibition of luciferase induction, and this inhibition by bortezomib persisted through 24 h after treatment. The bortezomib inhibition was statistically significant at all time points whether measured as absolute increases in luciferase signal or fold increase from baseline (not displayed).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

In vivo IVIS imaging of anesthetized, 3- to 6-mo-old Saa1-luc transgenic mice. Representative mice are shown for each group. Top panel, Male mice were pretreated with bortezomib (1 mg/kg i.v.) or saline 1 h before injection of LPS, (1 mg/kg i.p.). Middle panel, Female Saa1-luc transgenic mice were pretreated with bortezomib (1 mg/kg i.v.) or saline 1 h before intra-articular injection of zymosan (300 μg/knee joint) into the right rear knee joint and vehicle (5% glucose in saline) into the left rear knee joint. The zymosan-injected right knee area or the saline-injected left knee area was measured for total emission of photons. Bottom panel, Male Saa1-luc transgenic mice received oxazolone or vehicle topically to the shaved thorax area. The times shown represent hours after treatment. The color overlay on the image represents the photons per second emitted from the animal in accordance with the pseudo-color scale shown above the images. Red represents the highest number of photons per second, and blue represents the lowest number of photons per second.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

A, Change in liver region luciferase signal over time in male Saa1-luc transgenic mice pretreated with bortezomib (1 mg/kg i.v.; n = 10) or saline (n = 5) 1 h before injection of LPS (1 mg/kg i.p.). Five of the 10 bortezomib-pretreated mice were imaged at 24 h. Mean changes from baseline ± SE are presented. Nonparametric significance levels for the difference between treatment groups were determined by Mann-Whitney U test and are presented above the bars. B, Northern blot analysis of SAA1 mRNA. Induction of SAA1 mRNA expression was demonstrated in liver tissue from male and female mice treated with LPS (1 mg/kg i.p.) compared with that in control liver tissue. Pretreatment of the mice with bortezomib (1 mg/kg i.v.) inhibited LPS-induced SAA1 mRNA expression in the liver. Each RNA sample was prepared using tissues collected from three mice that were equally mixed. Equal loading was demonstrated by 18S rRNA.

Suppression of LPS-induced SAA1 expression by bortezomib occurred in multiple tissues

Because SAA1 promoter-driven luciferase induction by LPS occurred in several tissues, we investigated whether the suppression by bortezomib of luciferase occurred in all these tissues as well. Saa1-luc mice were pretreated with bortezomib (1 mg/kg i.v.) 1 h before LPS treatment. Positive control mice were treated with LPS alone, and negative control mice received saline. Selected organs were removed 4 h after LPS treatment and processed for luciferase activity; the data are presented in Table II⇓. Consistent with the findings shown in Table I⇑, the luciferase signal was significantly induced in all organs of LPS-treated mice compared with that in mice treated with saline. Mice pretreated with bortezomib showed significant inhibition of luciferase induction in their liver, lungs, kidney, and brains compared with mice that received only LPS. Compared with luciferase activity in the organs of control mice, LPS treatment caused a >9000-fold induction of liver luciferase activity, which was reduced by >80% by pretreatment with bortezomib. In lungs, kidney, and heart, the induction of luciferase expression was 200- to 400-fold by LPS, which was reduced by >60% with bortezomib (Table II⇓). Serum SAA1 levels 4 h after LPS treatment were significantly lower (p < 0.05) in mice pretreated with bortezomib (mean ± SE, 45 ± 7 μg/ml) than in those without pretreatment (1426 ± 46 μg/ml).

We analyzed the endogenous SAA1 mRNA expression in liver tissue, as shown in Fig. 4⇑B. In liver tissue of control mice, SAA1 mRNA expression was not detectable. An induction of SAA1 mRNA expression was observed in the livers of LPS-treated male and female mice. Induction of SAA1 mRNA expression by LPS was inhibited by bortezomib pretreatment. Thus, the luciferase activity correlated with the endogenous SAA1 mRNA expression in liver tissue. We also analyzed SAA1 mRNA expression in other nonhepatic tissues from control, LPS-treated, and bortezomib-pretreated mice. Due to the low sensitivity of the assay, we were unable to detect SAA1 message in any of the nonhepatic tissues (data not shown).

SAA1 promoter-driven luciferase expression is induced during acute arthritis, and the induction is suppressed by bortezomib

We next used the Saa1-luc model to study SAA1 promoter-driven luciferase expression during arthritis development. In this study, female Saa1-luc mice, 3–6 mo of age, were used for the induction of acute arthritis by intra-articular zymosan injection. As shown in Fig. 3⇑ (middle panel), injection of zymosan into the right knee joint of Saa1-luc mice induced luciferase expression at the knee joint. In addition, robust induction occurred in the liver. The induction of luciferase signal in knee joints and liver was detectable at 4 h, peaked at 7 h, and declined by 24 h after injection. At 24 and 120 h, the liver signal declined almost to the baseline level, whereas the signal from the knee joints remained induced. The control left knee that was injected with vehicle showed little change in luciferase activity. Pretreatment of the mice with bortezomib (1 mg/kg i.v.) significantly reduced luciferase induction in both liver and knee joint at 4 and 7 h (p < 0.05). Quantification of luciferase signal at the hepatic region and knee joint is presented in Fig. 5⇓. The zymosan-injected right knee joint showed 14-, 24-, 7-, and 6-fold inductions of luciferase signal at 4, 7, 24, and 120 h, respectively (Fig. 5⇓A). Bortezomib pretreatment moderately inhibited the signal from the right knee joint by ∼50% at 4 and 7 h and did not seem to affect the signal 24 and 120 h after zymosan injection (Fig. 5⇓A). The control left knee showed slight induction, possibly caused by a systemic response to zymosan administration (Fig. 5⇓B). In the hepatic region, the luciferase signal was highly induced 4 and 7 h after zymosan injection, whereas the pretreated mice showed <2-fold induction (Fig. 5⇓C). We also measured SAA1 protein production locally at the knee joint after intra-articular injection of either saline or zymosan. As shown in Fig. 5⇓D, the zymosan-injected knee joint had an SAA1 level of 0.42 μg/mg total protein, which is ∼12-fold higher than that in the saline-treated control knee joint.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

Change in luciferase signal over time in female Saa1-luc transgenic mice pretreated with bortezomib (1 mg/kg i.v.; n = 3) or saline (n = 3) 1 h before intra-articular injection of zymosan (300 μg/knee joint) into the right rear knee joint and vehicle (5% glucose in saline) into the left rear knee joint. Mean changes from baseline ± SE are presented. Two of the three bortezomib-pretreated animals were examined at 24 and 120 h. Nonparametric significance levels for the difference between treatment groups were determined by Mann-Whitney U test and are presented above the bars. Significant differences (p = 0.05) between the zymosan-injected (right) and vehicle-injected (left) rear knee joints are designated with asterisks. A, Measurement of luciferase signal from the zymosan-treated knee. B, Measurement of luciferase signal from the vehicle-treated knee. C, Measurement of luciferase signal from the hepatic region. D, Measurement of SAA1 protein production locally in the knee joint after intra-articular injection of either saline or zymosan. Mice (n = 3) were killed 6 h after the treatment. SAA1 measurement was performed by ELISA, as described in Materials and Methods.

SAA1 promoter-driven luciferase expression is induced during CHS reaction

We investigated the response of SAA1 promoter-driven luciferase expression during CHS reaction triggered by topical application of oxazolone. As shown in Fig. 3⇑ (lower panel), application of oxazolone triggered induction of luciferase expression in the treated area and hepatic region. The induction was first observed 4 h after treatment and peaked at 7 h in both the treated area and the liver. The induction was still high at 24 h. After that, the liver signal declined rapidly, whereas the signal from the treated area declined much more slowly and was still detected at 120 h. The vehicle-treated control mice showed a slight response. Quantification of the liver and oxazolone-treated skin area (Fig. 6⇓, A and B) showed that at the treated area, induction of luciferase in oxazolone-treated mice was significantly greater than that in mice treated with vehicle alone at 4, 7, 24, and 48 h after treatment. Significantly greater induction in the hepatic region occurred in oxazolone-treated mice relative to application of the vehicle 4 and 24 h after treatment (marginally greater at 7 h, p = 0.06). A 264-fold induction of luciferase activity was detected at 7 h in the hepatic region of oxazolone-treated mice. For the oxazolone-treated area, we detected a 39-fold induction of luciferase activity at the peak of 7 h, whereas the control vehicle-treated area showed a 2.5-fold induction. At 120 h in vehicle-treated mice, the signal from the liver and the vehicle-treated area declined to the basal level. In oxazolone-treated mice, the liver signal declined to the basal level, whereas the signal from the treated area was still 3-fold above the basal level at 120 h (Fig. 6⇓, A and B). Induction of SAA1 protein synthesis in the oxazolone-treated skin area was confirmed by ELISA. As shown in Fig. 6⇓C, oxazolone-treated skin tissue had a SAA1 level of 0.51 μg/mg total protein, which is ∼19-fold higher than that in vehicle-treated control skin tissue.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 6.

A and B, Change in luciferase signal over time in male Saa1-luc transgenic mice receiving oxazolone (n = 3) or vehicle (n = 3) topically to the shaved thorax area. Mean changes from baseline ± SE are presented. Nonparametric significance levels for the difference between treatment groups were determined by Mann-Whitney U test and are presented above the bars. Photons were collected from the treatment area and from the hepatic region. C, Measurement of SAA1 protein production locally in oxazolone-treated skin tissue. Mice (n = 3) were killed 6 h after the treatment. SAA1 measurement was performed by ELISA, as described in Materials and Methods. ∗, Significant difference (p = 0.05) between oxazolone-treated skin and vehicle-treated skin (control).