Role of matrix metalloproteinase-7 in the modulation of a Chlamydia trachomatis infection

Abstract

To determine the role of matrix metalloproteinase-7 (MMP-7) in the pathogenesis of chlamydial infection, C57BL/6 wild-type (WT) and MMP-7 knockout (KO) mice were infected intravaginally with Chlamydia trachomatis mouse pneumonitis (MoPn). Over a period of 6 weeks postinfection, various organs were cultured for C. trachomatis. Other infected animals were mated to assess their fertility status. No significant differences were observed between WT and KO mice in the number of animals with positive vaginal cultures, length of time of C. trachomatis shedding, or the number of C. trachomatis inclusion-forming units (IFU) recovered from their genital tracts. Likewise, the number of animals with hydrosalpinx, and the fertility rates and the number of embryos per mouse, were similar in WT and KO mice. Cultures from the spleen, lungs, kidneys and large intestine yielded similar numbers of IFU from WT and KO mice. However, the number of C. trachomatis IFU recovered from the small intestine of KO mice was significantly higher than that recovered from the small intestine of WT mice at 2 weeks postinfection. Because MMP-7 KO mice are deficient in active intestinal α-defensins, the results suggest that these components play a role in regulating colonization of the gastrointestinal tract by Chlamydia. By contrast, MMP-7 is dispensable in the progression and resolution of the genital tract infection.

Introduction

Matrix metalloproteinases (MMPs) comprise a family of structurally related zinc-dependent endopeptidases.^1,2 Traditionally, MMPs have primarily been associated with the catabolism of extracellular matrix (ECM) components, including those of basement membranes. However, recent work has revealed that these enzymes also recognize non-ECM substrates that are involved in regulating cellular homeostasis, tissue repair and inflammation.³ In addition, some MMPs may directly participate in various aspects of the immune response. For example, MMP-7, also called matrilysin, is abundantly expressed in mouse small intestinal Paneth cells, where the enzyme activates precursors of α-defensin antimicrobial peptides.^4,5 As prominent molecules of innate immunity, defensins are made not only by inflammatory cells, but also by a variety of mucosal epithelia.⁶ Similarly, in both rodents and humans, MMP-7 is constitutively produced by mucosal epithelium, especially ductal and glandular epithelium, including that of the respiratory, genitourinary and gastrointestinal tracts.^4,7–10

In the rodent female reproductive tract, MMP-7 is expressed in the cervix¹⁰ and localizes to epithelial cells of the uterus⁴ and oviducts (C. L. Wilson, unpublished observations). Basal levels of MMP-7 in the uterus and cervix fluctuate with the oestrus cycle,^10,11 as they do in the endometrium during the menstrual cycle in humans.⁷ In addition, in many epithelial cells, MMP-7 is up-regulated by infection and other types of injury.^12–14 Although the relevant substrate(s) for MMP-7 in both normal and infected mucosal epithelium have yet to be identified, we propose that, given the prominent induction of this enzyme in response to bacteria, the target molecules for MMP-7 are probably involved in host defence.

Infections caused by Chlamydia trachomatis are widespread throughout the world and cause a variety of diseases.^15–17 There are three biovars of C. trachomatis: the trachoma biovar mainly infect the epithelium of the eye, and the genitourinary, respiratory and gastrointestinal tracts in humans; the lymphogranuloma venerum biovar for the most part affects the reticuloendothelial system in humans; and the mouse pneumonitis (MoPn) biovar has been isolated only from mice inoculated with human lung tissues.^16,18 Most of the infections with the trachoma isolates are asymptomatic or produce acute symptoms, and the patients recover within a period of a few weeks.¹⁶ However, in certain individuals, C. trachomatis infections lead to long-term sequelae. For example, in regions with poor socioeconomic conditions, infections of the eye with this bacterium may result in trachoma, the most common cause of preventable blindness in the world.^15,16 Genital infections may also produce long-term sequelae, including infertility, ectopic pregnancy and chronic abdominal pain.^17,19

The mechanisms that lead to the development of long-term sequelae by C. trachomatis are poorly understood, but, in most instances, there is involvement of the ECM with production of scar tissue.^15–17 Thus, MMPs, and in particular MMP-7, could affect the outcome of an infection with C. trachomatis at several stages of the disease process:

• by playing a role in innate immunity, MMP-7 could participate in the control of this pathogen during the early stages of the infection;

• because of its ability to degrade components of the ECM, MMP-7 could modulate the outcome of the long-term sequelae resulting from a C. trachomatis infection; and/or

• MMP-7 could function in regulating the inflammatory response to Chlamydia infection of the upper genital tract because the enzyme has been shown to control chemokine mobilization and subsequent transepithelial movement of neutrophils in the injured lung.²⁰

To test these possibilities, we infected the genital tract of wild type (WT) and MMP-7-knockout (KO) mice with C. trachomatis and monitored the course of the disease. Inoculation of the genital tract of mice with the C. trachomatis MoPn serovar parallels the infection observed in humans, as shown by the development of salpingitis and infertility.^21,22

Materials and methods

Organisms

The C. trachomatis MoPn biovar (strain Nigg II; also called C. muridarum) was purchased from the American Type Culture Collection (Manassas, VA) and grown in HeLa-229 cells.²¹ Elementary bodies were isolated as described by Caldwell et al.²³ Organisms were frozen at −70° in buffer comprising 0·2 m sucrose, 20 mm sodium phosphate, pH 7·4, and 5 mm glutamic acid.

Infection with C. trachomatis

The seven- to eight-week-old C57BL/6 (H-2^d) WT and MMP-7 KO (backcrossed for 10 generations to C57BL/6) female mice used in these studies were bred at the Washington University School of Medicine (St Louis, MO). The animal protocol was approved by the University of California, Irvine, Animal Care and Use Committee. Lack of MMP-7 expression in the small intestine and reproductive tract of KO mice was confirmed (ref. ²⁴, and C. L. Wilson and L. Matrisian, unpublished observations). Mice were infected intravaginally with 10⁶ inclusion-forming units (IFU) of C. trachomatis MoPn.²² Vaginal swabs were collected and cultured at 7-day intervals for a period of 6 weeks postinfection (p.i.), as described previously.²¹ In addition, groups of three to five mice were killed at 5, 15 and 35 days p.i., and their organs were cultured for Chlamydia, as previously described.²¹ All experiments were repeated at least twice.

Humoral immune response to C. trachomatis

Serum samples and vaginal washes were collected at various intervals p.i., pooled for each group, and processed as previously described.²¹ In brief, 96-well plates were coated with 100 µl per well of purified C. trachomatis MoPn elementary bodies, at a concentration of 10 µg/ml of protein in phosphate-buffered saline (PBS). Subsequently, 100 µl of serum, or 50 µl of the vaginal wash, was added to each well in twofold serial dilutions. Samples were incubated at 37° for 1 hr and were then washed extensively with PBS. Horseradish peroxidase-conjugated goat anti-mouse immunoglobulin G (IgG), IgG1, IgG2a, immunoglobulin M (IgM) or immunoglobulin A (IgA) (Southern Biotechnology Associates Inc., Birmingham, AL) were added to the plates. Following incubation, the wells were washed and antibody binding was measured in an enzyme-linked immunosorbent assay (ELISA) reader (Bio-Rad Laboratories, Richmond, CA) using 2-2′-azinobis(3-ethylbenzthiazoline-6-sulphonic acid) as the substrate. The levels of antibody are expressed as the geometric mean titre.

Fertility studies

Six weeks after the intravaginal challenge, groups of four female mice were housed with a proven breeder male mouse for a maximum of 18 days, and the pregnancies were assessed by measuring the weight of each mouse.^21,22 Animals that had gained 5–10 g of weight by or before 18 days postmating were considered pregnant, were killed and the number of embryos in each uterine horn was counted. After the first mating, the female mice that did not gain weight were mated a second time with a male mouse that had successfully mated with another group of female mice and were monitored as described previously.²¹ All the animals that had not gained weight were killed 25 days from the start of the second mating. The number of animals with hydrosalpinx, and the number of embryos in each uterine horn, were counted when the mice were killed.

Statistics

Statistical analyses were performed using the statview software package on a Macintosh computer. The two-tailed unpaired Student's t-test, the Mann–Whitney U-test, and the Fisher's exact test were employed to determine the significance of differences between groups.

Results

Vaginal cultures for C. trachomatis MoPn

Mice were infected intravaginally with the C. trachomatis MoPn serovar and the course of the infection was followed by performing weekly vaginal cultures. As shown in Table 1, C. trachomatis was recovered for a period of 3 weeks in both the WT and the KO mice. At weeks 1 and 2 p.i., the number of WT and KO mice with positive cultures was similar. By contrast, a higher percentage of KO mice had positive vaginal cultures (39% versus 17% of WT) 3 weeks after infection, although this difference was not statistically significant (P > 0·05). Similarly, no significant differences were observed in the intensity of the infection between the two groups, as measured by the number of IFU recovered from the weekly vaginal cultures. By the 5th week after infection, 13% (three of 23) of the KO mice still had positive cultures, while none of the cultures from the WT mice were positive; again, this difference was not statistically significant (P > 0·05). All the animals had negative cultures 6 weeks after the vaginal infection.

Table 1.

Results of vaginal cultures in mice infected with Chlamydia trachomatis mouse pneumonitis (MoPn)

		Mice (%) with positive vaginal cultures, and median (range) no. of C. trachomatis MoPn IFU (×103) per mouse per week
		Week 1		Week 2		Week 3		Week 4		Week 5
Mice	No. of mice/group	IFU	%	IFU	%	IFU	%	IFU	%	IFU
MMP-7 WT	23	82·6	1·6 (0–110·6)	65·2	0·4 (0–74·3)	17·4	0 (0–2·7)	0	0 (0)	0	0 (0)	19 (82·6)
MMP-7 KO	23	91·3	3·2 (0–367·3)	65·2	0·3 (0–43·9)	39·1	0 (0–16·9)	0	0 (0)	13·0	0 (0–5·8)	21 (91·3)

IFU, inclusion-forming units; KO, knockout; MMP, matrix metalloproteinase; WT, wild type.

Organ culture for C. trachomatis MoPn

Groups of three to five mice were killed at 5, 15 and 35 days following vaginal infection, and several organs were harvested and cultured for Chlamydia. All WT and KO mice had positive cultures from the small intestines, and the majority of the animals also had positive cultures from the large intestine at the three time-points studied (Table 2). No significant differences in the number of C. trachomatis IFU recovered from the large intestine were observed between the WT and the KO mice (P > 0·05; Table 2). By contrast, cultures from the small intestine of KO mice had a greater number of C. trachomatis IFU than the cultures from WT animals, particularly on day 15 p.i. (P < 0·05). On day 5 p.i., both the median and range of recoverable IFU were higher from KO small intestine compared with WT small intestine. While the median value was the same for both genotypes at 35 days p.i., the upper end of the range was higher in KO versus WT small intestine. A few animals from both groups had positive cultures from the spleen, lungs and kidneys, but there were no statistically significant differences between the WT and the KO mice (Table 2). All cultures from the liver and heart were negative (data not shown).

Table 2.

Culture results from various organs of mice infected intravaginally with Chlamydia trachomatis mouse pneumonitis (MoPn)

		Mice (%) with positive cultures, and median (range) no. of C. trachomatis MoPn IFU recovered from:
		Small intestine		Large intestine		Spleen		Lungs		Kidneys
Mice	Days p.i.
IFU (× 104)	%	IFU (× 104)	%	IFU	%	IFU	%	IFU
MMP-7 WT	5	100	15·5 (0·05–66·6)	67	23·0 (0–37·5)	0	0	0	0	0	0 (0–100)
MMP-7 KO	5	100	41·6 (2·9–263·0)	100	25·0 (1·5–136·0)	40	0 (0–75)	0	0	0	0
MMP-7 WT	15	100	0·5 (0·005–5·4) *	100	18·7 (0·1–34·9)	20	0 (0–4)	20	0 (0–50)	0	0
MMP-7 KO	15	100	18·2 (4·9–3240·0)	100	7·8 (4·1–900·0)	0	0	0	0	20	0 (0–50)
MMP-7 WT	35	100	0·5 (0·2–1·0)	100	240·0 (70·5–630·0)	0	0	0	0	0	0
MMP-7 KO	35	100	0·5 (0·1–1251·0)	100	20·5 (20·0–348·0)	0	0	0	0	0	0

IFU, inclusion-forming units; KO, knockout; MMP, matrix metalloproteinase; p.i. postinfection; WT, wild type.

^*P < 0·05 (Mann–Whitney U-test).

Antibody response

To assess the role that MMP-7 may play in the modulation of the immune response to a C. trachomatis infection, serum and vaginal samples were collected over the 6-week period of the study, and Chlamydia-specific antibody titres were determined by ELISA. IgG antibodies were detected in the serum of both the KO and WT mice at 2 weeks p.i. (Table 3). The antibody titres progressively increased and plateaued during the 5th week p.i. Both WT and KO mice mounted predominantly a T helper 1 (Th1) response, as shown by the IgG2a/IgG1 ratio. Similarly, IgG and IgA Chlamydia-specific antibody levels were detected in vaginal washes from both WT and KO mice. No significant differences were observed in the serum or vaginal antibody titres between the WT and the KO mice.

Table 3.

Serum and vaginal antibody enzyme-linked immunosorbent assay (ELISA) titres in mice following an intravaginal infection with Chlamydia trachomatis mouse pneumonitis (MoPn)

	Serum						Vagina
	MMP-7 KO mice			WT mice			MMP-7 KO		WT mice
Days p.i.	IgG1	IgG2a	IgG	IgG1	IgG2a	IgG	IgA	IgG	IgA
″0	< 100	< 100	< 100	< 100	< 100	< 100	< 10	< 10	< 10	< 10
″9	400	< 100	100	200	< 100	< 100	< 10	< 10	< 10	< 10
″15	6400	< 100	400	12 800	< 100	1600	20	40	10	10
″35	102 400	3200	51 200	51 200	800	12 800	80	160	40	80
″44	102 400	3200	25 600	51 200	800	6400	80	80	40	160

KO, knockout; MMP, matrix metalloproteinase; WT, wild type.

Fertility studies

Six weeks after female mice were inoculated in the vagina with C. trachomatis, male mice were housed in the same cage and the resulting pregnancies were followed over a course of two mating cycles. As shown in Table 4, infected WT and KO mice exhibited similar levels of infertility. The number of embryos per mouse was also comparable for both groups. Moreover, no significant differences were noted between the WT and the KO mice in the number of animals that had hydrosalpinx (39% versus 22%).

Table 4.

Results of fertility studies in mice infected in the vagina with Chlamydia trachomatis

Mice	No. of mice with uni- or bi-lateral infertility/ total no. of mice (%)	No. of mice with hydrosalpinx/ total no. of mice (%)	Mean 1 no. of embryos in both uterine horns/mouse
MMP-7 WT	15/23 (65·2)	9/23 (39·1)	3·5 ± 0·8
MMP KO	14/23 (60·9)	5/23 (21·7)	3·3 ± 0·6

KO, knockout; MMP, matrix metalloproteinase; WT, wild type.

¹Mean ± 1 standard error.

Discussion

Using a model of vaginal infection of C57BL/6 WT and MMP-7 KO mice with C. trachomatis MoPn, we have demonstrated that MMP-7 appears to play a limited role in the pathogenesis of the disease in the genital tract. Neither the course of the infection nor the long-term sequelae resulting from a genital chlamydial infection were different in MMP-7 KO mice when compared with WT animals. Inoculation of mice by any mucosal route with the C. trachomatis MoPn serovar results in a disseminated infection that affects multiple organs, including other mucosal sites, the spleen, liver and kidneys.²⁵ While a significant inflammatory response occurs at most sites of infection, leading, in some instances, to long-term sequelae, in the intestinal mucosa there is no tissue pathology.²⁶ Here, infection of the genital tract with C. trachomatis MoPn resulted in dissemination of this bacterium to other organs. While no difference was noted in the numbers of bacteria recovered from the kidney, lung, spleen and large intestine of WT versus KO mice, greater numbers of infectious Chlamydia were isolated from the small intestine of the KO animals.

MMPs have been implicated in the pathogenesis of chlamydial infections because they are considered to be classical matrix remodelling enzymes, and the long-term sequelae of these infections are primarily the result of scar tissue formation. 15–17 The expression of several MMPs has been investigated in both human specimens and in experimental models of chlamydial infection. For example, Abu El-Asrar et al.^27,28 noted increased levels of MMP-9 in monocyte/macrophages and in polymorphonuclear leucocytes in conjunctival biopsies from patients with active trachoma as compared with controls. Zymographic analysis also showed high levels of MMP-9 in the conjunctiva from trachoma specimens in these studies and, more recently, MMP-9 mRNA was demonstrated to be highly up-regulated during active trachoma.²⁹ In addition, MMP-9, as well as MMP-12, was detected in the upper genital tract in Chlamydia-infected female C3H/HeN mice, a strain highly susceptible to the bacterium.^22,30 MMPs may be produced as part of the host attempt to eradicate the infection, but enhanced expression of these enzymes could also lead to matrix remodeling and scar formation.

In the studies described in this report, we sought to determine whether MMP-7 functions in modulating the course of a Chlamydia infection. We focused on MMP-7 for three key reasons.

1. The enzyme is expressed in the cervix and in epithelial cells of the uterus⁴ and oviduct (C. L. Wilson, unpublished observations) and is potently induced in epithelial cells exposed to bacteria.^12,13,14 Moreover, the epithelium is the primary target of Chlamydia.^15,16

2. MMP-7 proteolytically modulates a broad range of both matrix and non-matrix substrates; thus, the enzyme could be involved at several stages of the infection and in the host response.

3. MMP-7 has already been shown to cleave and activate antimicrobial peptides in mouse small intestinal Paneth cells, indicating that the enzyme has a role in gut innate immunity and could have a similar function in other epithelial tissues.⁵

In experimental infections of the mouse reproductive tract, the oestrus cycle in the animals is often synchronized by the administration of progesterone prior to inoculation with C. trachomatis. We elected not to use progesterone pretreatment in our studies because this compound not only is a strong immune modulator, but it also represses expression of MMPs, especially MMP-7, in the endometrium (reviewed in ref. 31).

We infected both WT and MMP-7 KO female mice with C. trachomatis MoPn to compare disease parameters between the two genotypes, including the antibody response, the effects on vaginal bacterial load and the long-term sequelae, including infertility and the formation of hydrosalpinx. None of these parameters differed significantly in MMP-7 KO mice when compared with WT animals. These findings suggest that this MMP is dispensable in the pathogenesis of the disease. It is possible that other MMPs are able to compensate for the lack of MMP-7 in this infection model. For example, using human fallopian tube organ cultures, Ault et al.³² found increased expression, as well as activation, of both MMP-2 and -9 at 2 and 3 days following an in vitro infection with the human serovar E of C. trachomatis. Using immunohistochemistry and in situ zymography, these investigators found that while MMP-9 was localized diffusely throughout the stroma, MMP-2 was present on the epithelial cells of the infected tubes. Hence, MMP-2 is a candidate compensatory molecule, although MMPs are not always functionally interchangeable. In addition, although mice lacking MMP-7 showed no differences from WT animals in effects on the course of the infection and long-term sequelae, there may be subtle differences between the two genotypes in the inflammatory response, especially in the upper genital tract, that we have not yet uncovered.

In addition to the eye and the genital tract, the gastrointestinal tract is a site that can harbor Chlamydia. In adults, particularly homosexuals, Chlamydia can often be recovered from the rectum.³³ Although in a large number of cases colonization by Chlamydia of the gastrointestinal tract of humans or animals is asymptomatic, a variety of clinical manifestations, ranging from mild diarrhoea to severe symptomatology, have been described. For example, a possible role for chlamydial infections has been postulated in the aetiopathogenesis of inflammatory bowel disease and, in particular, Crohn's disease.^33,34 In addition, the LGV strains of C. trachomatis can produce very severe long-term sequelae with scarring and stenosis of the gastrointestinal tract.¹⁵

Following the vaginal infection, we detected no consistent differences in the numbers of Chlamydia or patterns of colonization in the spleen, kidneys, lungs and large intestine of WT and KO mice. However, the small intestine was a notable exception. Here, higher levels of Chlamydia were recovered from the KO mice than the WT mice, particularly 2 weeks following intravaginal infection. Although we cannot rule out the possibility that MMP-7 contributes in some way to the adaptive immune response, we speculate that the difference in recoverable bacteria may be caused by the deficiency in activated defensins, or cryptdins, in MMP-7 KO mice.⁵ MMP-7 is produced constitutively in the Paneth cells of the mouse small intestine and mediates the processing and activation of cryptdins in these cells.^5,35 Indeed, ≈ 70% of the antimicrobial activity of the Paneth cell is attributable to defensins.³⁶ Oral infection of mice with Escherichia coli was found to result in higher numbers of bacteria recoverable from the small intestine of KO versus WT animals,⁵ akin to the observations reported here for Chlamydia. In addition, it has been shown that KO mice succumb more rapidly and at a lower oral dose of virulent Salmonella enterica serovar typhimurium than WT mice.⁵ In the case of C. trachomatis, several antimicrobial peptides have been tested in vitro for their antichlamydial activity.^37,38 Among them, protegrin-1 had the strongest antichlamydial activity against C. trachomatis MoPn. Although cryptdin-4, one of the six cryptdin peptides that have been characterized, did not significantly affect the in vitro growth of MoPn,³⁸ it may be that a combination of cryptdin peptides is effective against this organism in vivo.

In humans, the mechanism of enteric α-defensin activation differs from that in mice. The precursor of the human Paneth cell α-defensin, HD-5, is proteolytically processed by trypsin, not MMP-7.³⁹ Nevertheless, our studies, using MMP-7-deficient (and thus defensin-deficient) mice, indicate that these components of the innate host defence may play a role in regulating gastrointestinal colonization by Chlamydia. The recent development of transgenic mice expressing HD-5 in Paneth cells⁴⁰ provides an excellent opportunity for testing the possibility that defensins also modulate intestinal Chlamydia in humans. In this model, we predict that HD-5 transgenics would clear the bacteria from the small intestine better than WT mice.

In conclusion, we have shown that vaginal infection of C57BL/6 WT and MMP-7 KO mice with C. trachomatis does not result in any significant differences in the course of the acute genital infection and the long-term sequelae. However, a higher load of infectious Chlamydia was recovered from the small intestine of the KO mice than in the WT mice, indicating that MMP-7 plays a role in the local control of this infection in mice, potentially through proteolytic activation of enteric α-defensins. The specific mechanisms involved in the control of Chlamydia in the gastrointestinal tract are currently under investigation in our laboratories.

Abbreviations

ECM

extracellular matrix

ELISA

enzyme-linked immunosorbent assay

IFU

inclusion-forming units

IgA

immunoglobulin A

IgG

immunoglobulin G

IgM

immunoglobulin M

MMP

matrix metalloproteinase

MoPn

mouse pneumonitis

KO

knockout

PBS

phosphate-buffered saline

p.i.

postinfection

Th1

T helper 1

WT

wild type