TIMP3 Regulates Mammary Epithelial Apoptosis with
Immune Cell Recruitment Through Differential TNF
Dependence
Carlo V. Hojilla, Hartland W. Jackson, Rama Khokha*
Department of Medical Biophysics, Ontario Cancer Institute, University of Toronto, Toronto, Ontario, Canada
Abstract
Post-lactation mammary involution is a homeostatic process requiring epithelial apoptosis and clearance. Given that the
deficiency of the extracellular metalloproteinase inhibitor TIMP3 impacts epithelial apoptosis and heightens inflammatory
response, we investigated whether TIMP3 regulates these distinct processes during the phases of mammary gland
involution in the mouse. Here we show that TIMP3 deficiency leads to TNF dysregulation, earlier caspase activation and
onset of mitochondrial apoptosis. This accelerated first phase of involution includes faster loss of initiating signals (STAT3
activation; TGFb3) concurrent with immediate luminal deconstruction through E-cadherin fragmentation. Epithelial
apoptosis is followed by accelerated adipogenesis and a greater macrophage and T-cell infiltration in Timp3
2/2
involuting
glands. Crossing in Tnf deficiency abrogates caspase 3 activation, but heightens macrophage and T-cell influx into Timp3
2/2
glands. The data indicate that TIMP3 differentially impacts apoptosis and inflammatory cell influx, based on involvement of
TNF, during the process of mammary involution. An understanding of the molecular factors and wound healing
microenvironment of the postpartum mammary gland may have implications for understanding pregnancy-associated
breast cancer risk.
Citation: Hojilla CV, Jackson HW, Khokha R (2011) TIMP3 Regulates Mammary Epithelial Apoptosis with Immune Cell Recruitment Through Differential TNF
Dependence. PLoS ONE 6(10): e26718. doi:10.1371/journal.pone.0026718
Editor: John P. Lydon, Baylor College of Medicine, United States of America
Received September 27, 2011; Accepted October 3, 2011; Published October 28, 2011
Copyright: ß 2011 Hojilla et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: Dr. Hojilla was supported by a Natural Sciences and Engineering Research Council of Canada studentship and a National Cancer Institute of Canada
Terry Fox Studentship. Dr. Jackson is supported by a Canadian Breast Cancer Foundation Fellowship. This work has been funded by Canadian Breast Cancer
Research Alliance grant #16411 to Dr. Khoka. The funders had no role in the study design, data collection and analysis, decision to publish, or preparation of the
manuscript.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: rkhokha@uhnres.utoronto.ca.
Introduction
Tissue involution is a postnatal process that offers a unique
window to study physiological cell death. Involution occurs in the
thymus [1], prostate [2], uterus and mammary gland [3]. These
tissues are subject to physiological cues that instruct them to
undergo apoptosis to maintain cell number and homeostasis. In
humans, the mammary gland undergoes extensive involution after
the cessation of lactation [4], and this process has been widely
studied in murine models. Post-lactational mammary gland
involution serves to remodel the highly structured secretory gland
into one that resembles the virgin state so that the differentiation
program initiated by gestation may begin anew [5,6]. Mammary
involution depends on epithelial apoptosis in which epithelial
lumens collapse and lobulo-alveolar structures are deleted with
rapid elimination of up to 90% of the epithelium [7]. This process
occurs in parallel with adipogenesis [8], and microarray expression
analyses have suggested the involvement of inflammatory
processes in mammary gland involution [9,10]. Yet, factors that
control these systems in parallel remain to be elucidated.
Mammary gland involution in the mouse has been described as
a two-step process defined by the reversibility of involution
[11,12]. The first stage is marked by increased pro-apoptotic
factors and can be reversed by re-induction of suckling. Death
receptor-mediated and mitochondrial apoptosis signaling path-
ways are employed by mammary epithelial cells during involution.
TNF and p55-TNFRI mRNA transcript levels decrease, while
p75-TNFRII levels increase in the involuting mammary glands of
rats [13]. Fas and FasL levels in glandular epithelia increase
immediately upon pup withdrawal. As such, Fas-deficient lpr mice
and FasL-deficient gld mice do not show apoptotic cells until the
later stages of involution [14]. Specifically, TGFb3, STAT3 and
their downstream targets have been identified as initiators of the
first phase of involution [15,16,17]. Mammary gland involution is
also associated with cytochrome c release from the mitochondria,
as well as the processing of initiator (caspases 8, 9) and executioner
caspases (caspases 3, 7) which generally peak at 3 days after
cessation of suckling in wild-type mice [14,18].
The second irreversible phase of involution occurs around day 3
and involves widespread apoptosis and tissue remodeling, with
extracellular matrix turnover and structural reorganization of the
gland [19]. Matrix metalloproteinases (MMPs) and their tissue
inhibitors (TIMPs) regulate matrix turnover during pubertal
morphogenesis [20,21], gestation, lactation [22], and involution
[23,24,25]. While the first phase of involution is considered to be
protease-independent [11], the second protease-dependent stage
has decreased levels of TIMP and activated MMPs which initiate
the deconstruction of the surrounding stroma therefore facilitating
gland remodeling. At the same time, MMP activity has also been
proposed to influence adipogenesis [8]. Synchronization of
PLoS ONE | www.plosone.org 1 October 2011 | Volume 6 | Issue 10 | e26718
multiple compartments is essential for maintaining the functional
integrity of the mammary gland during the phases of involution.
Interestingly, microarray and histological studies have revealed
that inflammatory mediators are present during mammary gland
involution [9,10]. These arrays suggest the death receptor family
involvement in the first phase of involution and the involvement of
monocyte and lymphoid cytokines during the second phase [26].
Macrophages in particular have been found in the involuting
glands of several species, and respond to the chemotactic signals
provided by the remodeling matrix [27]. The involuting
mammary gland microenvironment has been likened to wound
healing and cancer environments [28]. Macrophages are known to
play a role in involution, the wound healing response, and are also
implicated in promoting breast cancer metastasis [27,29]. Given
that pregnancy associated breast cancers (PABC) have a poor
patient outcome [30,31,32], a better definition of molecular factors
that differentially regulate apoptosis from inflammatory cell
recruitment will enhance our understanding of the processes
underscoring physiological mammary involution and mammary
cancer progression.
Both gain- and loss-of-function studies underscore the complex
apoptotic role of TIMP3 [33]. TIMP3 is a negative regulator of
inflammation in specific tissues [34,35,36] and in systemic
responses to endotoxin [37]. TIMP3 deficient mice have
accelerated post-lactation involution and fail to reinitiate lactation
after 48hrs [23]. In this study, we ask whether TIMP3 functions to
integrate the distinct programs of apoptosis and inflammation
during homeostatic remodeling of post-lactational mammary
involution. We identify the molecular and cellular events
underlying the accelerated mammary gland involution in
Timp3
2/2
mice. TIMP3 deficiency leads to a loss of the survival
signal provided by E-cadherin-mediated cell to cell contact as well
as an accelerated first phase of involution. A far greater and
sustained influx of macrophages also occurs, which is coupled with
an infiltration of T-cells in Timp3
2/2
involuting glands. Combin-
ing the TIMP3 and TNF deficiencies, we further dissect the TNF
dependency of these molecular and cellular events. Our results
show that TIMP3 is a critical physiological regulator for both
apoptosis and inflammatory cell recruitment during murine
mammary gland involution.
Results
Higher TNF levels and accelerated apoptosis during
mammary involution in Timp3b mice
TIMP3 is the only TIMP capable of inhibiting the TNF alpha
converting enzyme (TACE/ADAM17) [38,39]. Timp3
2/2
mice
have increased TNF levels and signaling owing to increased
TACE activity in several systems [34,35,37]. We measured
mammary TNF levels, which were detectable at the end of
lactation, and again on day 4 of involution (4i) in wild-type mice.
Timp3
2/2
mammary gland also displayed two distinct periods of
TNF increase. It was 4-fold higher than wild-type at 10L,
reappeared earlier and stayed higher than in wild-type tissue
(Fig. 1A). Fas and FasL levels did not differ between wild-type and
Timp3
2/2
glands throughout involution (data not shown). Death
receptor-mediated apoptosis induced by TNF occurs through
activation of the initiator caspase, caspase 8 [40]. Consistent with
the elevated TNF levels in Timp3
2/2
glands, the cleaved form of
caspase 8 was detectable at 10L and elevated earlier than wild-type
(Fig. 1B).
We next examined the executioner arm of mitochondrial
apoptosis, a multiprotein complex known as the apoptosome,
which is comprised of cytochrome c, Apaf-1, and procaspase 9
[41,42,43]. These components showed an earlier temporal shift
upon the loss of TIMP3 (Fig. 1B). Notably, Apaf-1 was present on
1i and 2i coinciding with the timing of maximal epithelial
apoptosis in Timp3
2/2
mammary glands (Fata et al. 2001), but
absent in wild-type tissue. Caspases 3 and 9 were also clearly
activated earlier in TIMP3 null tissue as indicated by the presence
of their cleaved forms (Fig. 1B). The loss of mitochondrial
membrane integrity results in the release of cytochrome c into the
cytoplasm [44]. Immunofluorescence showed mitochondrial
cytochrome c release within luminal epithelial cells at 1i in
Timp3
2/2
glands, but only at 3i in wild-type (Fig. 1C, arrows). The
number of epithelial cells shed into lumens (Fig. 1C, arrowheads)
with cytochrome c release was present in greater numbers in
Timp3
2/2
glands compared to wild-type at 1i. We also observed
apoptosis-inducing factor (AIF) expression to be earlier than wild-
type in Timp3
2/2
involuting glands (Fig. 1B). The levels of second
mitochondria-derived activator of caspase (SMAC), a mitochon-
drial-derived inhibitor of IAPs (inhibitors of apoptosis) [45],
decreased sooner in Timp3
2/2
involuting glands (Fig. 1B). Thus,
an earlier progression to apoptosis was indicated by the rapid
appearance of several caspases and the loss of mitochondrial
integrity in Timp3
2/2
involuting mammary glands.
Mapping of first phase of involution and adipogenic
signals
Murine mammary gland involution occurs in two phases where
involution is reversible in the first 48 hours. Timp3 null mice fail to
re-establish lactation after 48 hours of involution [23]. STAT3 is
known to be involved in this early phase and STAT3 loss, or loss of
its upstream activators, abrogates the first phase and delays the
onset of involution [17,46,47]. We observed that total STAT3 was
upregulated earlier in Timp3
2/2
mammary gland. STAT3 was
phosphorylated immediately upon induction of involution in both
wild-type and Timp3
2/2
mammary glands, but activation of
STAT3 was turned off earlier in Timp3b and not seen past 3i
(Fig. 2). Milk stasis during involution also rapidly induces TGFb-3
levels in the mammary gland to induce apoptosis [15]. We
observed that Tgfb-3 expression lasted until 2i in wild-type glands,
but was undetectable after 1i in Timp3
2/2
glands (Fig. 2)
consistent with the rapid completion of the first phase of
involution.
In mammary involution, adipogenesis is important for the re-
establishment of the pre-pregnant state filling out the space left
behind by epithelial apoptosis [8]. We therefore examined markers
of adipocyte differentiation: CEBP-b for early [48] and PPAR-c
for late adipocyte differentiation, and ALBP as a marker for
adipocytes [49]. Generally, an earlier and increased levels of
PPARc and ALBP coincident with decreased CEBPb levels
indicated faster onset of the adipogenic program (Fig. 2).
E-cadherin breakdown compromises epithelial
architecture in Timp3
2/2
mammary glands
We determined whether TIMP3 loss affected E-cadherin-
mediated cell-cell contact since this adhesion molecule is processed
by MMPs [50,51], and we earlier reported an aberrant MMP2
activation during mammary involution in Timp3
2/2
mice [23]. E-
cadherin breakdown was evident by 2i in wild-type mammary
gland. On the other hand, an immediate and far greater
fragmentation of E-cadherin occurred at 1i in Timp3
2/2
mammary tissue (Fig. 3), which was even present at day 10 of
lactation (10L), suggesting that this tissue may have compromised
epithelial cell-cell contact during lactation. Also notable was the
overall elevated level of E-cadherin in Timp3
2/2
glands. There
TIMP3 Regulation of Mammary Involution
PLoS ONE | www.plosone.org 2 October 2011 | Volume 6 | Issue 10 | e26718
Figure 1. Increased TNF levels and apoptotic signaling in
Timp3
b involuting glands. (A) TNF levels as assessed by ELISA at different
involution timepoints starting from day 10L until day 7i revealed higher levels of TNF levels in Timp3
2/2
glands (n = 3) compared to wild-type (WT,
n = 3) both prior to the onset of involution and as an earlier rebound. (B) Representative western blot analyses of lysates from different involution
timepoints probed for caspase 8 (Casp 8), Apaf-1, cytochrome c (Cyt c), caspase 9 (Casp 9; top band represents pro-caspase 9, bottom band
represents cleaved caspase 9; ND = not determined for 7i), cleaved caspase 3 (Casp 3), Apoptosis-inducing factor (AIF), and Second mitochondria-
TIMP3 Regulation of Mammary Involution
PLoS ONE | www.plosone.org 3 October 2011 | Volume 6 | Issue 10 | e26718
was a corresponding disappearance of b-catenin by 4i in Timp3
2/2
involuting glands (Fig. 3) in contrast to wild-type tissue, suggesting
the earlier destabilization of E-cadherin-mediated adherens junc-
tions consistent with compromised cell-cell contact in this tissue.
Altogether the data illustrated in Figures 1, 2, and 3 provide a map
of key molecular signals involved in physiological apoptosis and
adipogenesis in the mammary gland, where TIMP3 deficiency leads
to an earlier deconstruction and remodeling of the epithelial
compartment.
Re-establishment of survival signaling by recombinant
TIMP3
Next we asked whether TIMP3 plays a direct role in inhibiting
E-cadherin breakdown and TNF bioactivity during mammary
involution. TIMP3-containing slow-release Elvax pellets were
implanted distal to the lymph node in the fourth inguinal
mammary fat pad at the onset of involution and tissue immediate
to the pellet was collected at 3i. Reconstitution of TIMP3 rescued
E-cadherin fragmentation in Timp3
2/2
mice (Fig. 4A). The
increased TNF level in Timp3
2/2
glands was also significantly
reduced (P,0.01) by this manipulation (Fig. 4B). Furthermore,
activation of caspase 8 was reduced in TIMP3-reconstituted null
tissue (Fig. 4C). These data show that TIMP3 protein could rescue
the alterations associated with accelerated mammary involution in
Timp3
2/2
mice.
Timp3 deficiency initiates inflammatory and immune cell
recruitment in involuting glands
Historically, studies in dairy animals (cow, sheep and goat) have
described neutrophil and macrophage infiltration during the
process of mammary gland involution [52,53,54]. Microarray
studies in the mouse also support an inflammatory component to
involution [9,10]. Since we previously identified TIMP3 as a
negative regulator of inflammation in several systems [34,55], we
assessed the infiltration of inflammatory cells into involuting
mammary glands. Neutrophil influx occurred during involution
and was comparable between wild-type and Timp3
2/2
involuting
glands (data not shown). In contrast, macrophage numbers were
comparable during lactation, but increased up to 4-fold (P,0.01)
in Timp3
2/2
involuting glands at early stage involution (1i and 2i)
when compared to wild-type glands, and rapidly declined
thereafter at 3i (Fig. 5A). The maximal macrophage influx in
wild-type glands was at 3i, however these did not reach levels seen
in Timp3
2/2
glands. This represented an abrupt and immediate
initiation of macrophage infiltration in mammary glands lacking
TIMP3.
The presence of lymphocytes has been indicated in the
involuting mammary glands of ruminants [56,57], but has yet to
be studied in the murine model. Given that lymphocytic infiltrates
were found in aged Timp3
2/2
livers [34], we measured the
infiltration of B and T-cells in involuting mammary glands. CD3-
positive T-cells were present in alveolar and ductal epithelium,
including interalveolar and periductal areas, at low levels in wild-
type but were elevated in timp3
2/2
glands throughout involution.
The number of T-cells in involuting Timp3-deficient glands peaked
at 3i and was sustained until 5i. The wild-type tissue displayed
similar kinetics but a smaller increase of these cells. The T-cell
infiltration exhibited different kinetics compared to the macro-
phage influx, temporally occurring subsequent to macrophage
(Fig. 5B). B cells were very sparse in the mammary gland and
remained comparable between wild-type and Timp3
2/2
glands, as
assessed by anti-B220 immunohistochemistry (data not shown).
The role of TNF in mammary involution of the Timp3
deficient gland
TIMP3 regulates the bioactivity of the inflammatory cytokine
TNF since it is the physiological inhibitor of the TNF sheddase
TACE. We have shown TIMP3 to be a critical regulator of
physiological systems that depend on TNF such as liver
regeneration and apoptosis [34,58] and the innate immune
response to endotoxins [36,37]. We therefore asked whether
increased apoptosis and inflammatory reaction during mammary
involution in Timp3
2/2
mice was dependent on increased TNF
Figure 2. Mapping of priming and adipogenic signals in
Timp3
2
/
2
mammary involution. Representative western blot analyses of lysates
from wild-type (WT) and Timp3
2/2
mammary glands at different involution timepoints starting from day 10L until day 7i probed for phosphorylated
STAT3 (P-STAT3), total STAT3, TGF-b3, and the adipogenic markers: CEBPb, PPAR-c, and ALBP.
doi:10.1371/journal.pone.0026718.g002
derived activator of caspase (SMAC). (C) Immunofluorescence and co-localization of mitochondria (MitoTracker) and cytochrome c revealed earlier
release of cytochrome c from the mitochondria (red signal, arrows) in Timp3
2/2
luminal cells, as well as cells that have detached and were shed into
the lumen by 1i, than in wild-type at 3i. Bar graphs are expressed as mean 6 SEM (* = P,0.02). Scale bar = 30
mm.
doi:10.1371/journal.pone.0026718.g001
TIMP3 Regulation of Mammary Involution
PLoS ONE | www.plosone.org 4 October 2011 | Volume 6 | Issue 10 | e26718
bioactivity, by combining the Timp3 and Tnf deficiencies. For these
experiments we used Timp3 mice that we backcrossed seven times
into the C57BL/6 background to match the Tnf deficient mice,
whereas the data in Figures 1, 2, 3, 4, and 5 were generated using
pure FVB/N Timp3
2/2
mice. Activation of caspase 3 is a measure
of apoptosis downstream of death receptor signaling by TNF, and
its positivity has been previously reported in regressing mammary
glands [59]. The percentage of mammary epithelial cells that
stained positive for active caspase 3 as quantified by histomor-
phometry of single (Timp3
2/2
or Tnf
2/2
) and compound
(Timp3
2/2
/Tnf
2/2
) knockouts in the C57BL/6 background at
1i and 3i (Fig. 6A). The higher number of activated caspase 3-
positive cells in Timp3
2/2
glands compared to wild-type confirmed
greater apoptosis in C57BL/6 Timp3
2/2
mice at 1i similar to
FVB/N Timp3
2/2
strain [23]. Also as expected, the apoptosis
index was higher in wild-type at 3i compared to Timp3
2/2
mice,
which coincide with the expected peak of epithelial apoptosis in
involuting wild-type glands. Deletion of Tnf in Timp3 null mice
resulted in significantly reduced cleaved caspase 3 positive cells
compared to Timp3
2/2
at 1i, showing that the lack of TNF
rescued the accelerated apoptosis in Timp3
2/2
glands (P,0.05;
Fig. 6A). A control group with TNF deletion alone also showed
reduced numbers of epithelial cells with activated caspase 3 at 3i,
also confirming the role of the TNF pathway during normal
mammary involution.
We next examined whether the greater inflammatory cell influx
seen during mammary involution of Timp3
2/2
mice depended on
TNF (Fig. 5, 6B, C). Intriguingly at 1i, the number of F4/80-positive
macrophages was increased by at least 4-fold in Timp3
2/2
/Tnf
2/2
when compared to Tnf
2/2
glands, and up to 8-fold (P,0.01) when
compared to Timp3
2/2
glands (Fig. 6B, 5C). At 3i, the overall
numbers of macrophages subsided and became comparable
between single and compound knockout groups, but still remained
higher than that in wild-type glands (Fig. 6B). Examination of CD3-
positive T-cells revealed a significant 3-fold increase (P,0.05) at 1i
following compound removal of TIMP3 and TNF (Fig. 6C). This
trend of increased T-cell recruitment was maintained at 3i in
Timp3
2/2
/Tnf
2/2
involuting glands (P,0.05; Fig. 5C). Contrary to
our expectation the inflammatory and immune cell influx was not
ameliorated by adding Tnf deficiency to Timp3-deficient mice. This
series of experiments indicated that TNF is required for accelerated
mammary epithelial apoptosis but not immune cell influx of Timp3
null involuting glands.
Discussion
The rapid mammary involution established in the Timp3 null
mouse provides a model to study the different phases and cellular
compartments of this remodeling tissue. We analyzed several key
molecules essential for epithelial cell survival and explored
inflammatory cells during mammary gland regression, as tempo-
rally mapped in Figure 7. The loss of the extracellular
metalloproteinase inhibitor, TIMP3, accelerated the first phase
of involution leading to an early onset of the irreversible second
phase [23]. We observed extensive E-cadherin fragmentation and
an immediate onset of TNF-induced apoptosis coinciding with
Figure 4. Inhibition of E-cadherin fragmentation and TNF
signaling by recombinant TIMP3. Pellets containing vehicle control
(ctrl) or recombinant human TIMP3 (rT3) were implanted at the
beginning of involution and tissues were harvested at day 3 of
involution. (A) Pellets containing recombinant TIMP3 blocked E-
cadherin fragmentation at 3i in Timp3
2/2
(n = 3). (B) The increase in
TNF levels at 3i in Timp3
2/2
tissue glands was reversed by recombinant
TIMP3 as measured by ELISA. (C) Caspase 8 activation at 3i was only
abrogated in Timp3
2/2
regressed tissue that was reconstituted with
recombinant TIMP3. Bar graphs are expressed as mean 6 SEM
(* = P,0.01; n = 3).
doi:10.1371/journal.pone.0026718.g004
Figure 3. Fragmentation of E-cadherin and disruption of adherens junctions. (A) Representative western blot analyses of lysates from wild-
type (WT) and Timp3
2/2
mammary glands at different involution timepoints starting from day 10 of lactation (10L) until day 7 of involution (7i)
probed for E-cadherin and b-catenin, showing early processing of E-cadhern and b-catenin in Timp3
2/2
mammary glands.
doi:10.1371/journal.pone.0026718.g003
TIMP3 Regulation of Mammary Involution
PLoS ONE | www.plosone.org 5 October 2011 | Volume 6 | Issue 10 | e26718
pSTAT3 activation and TGFb3 expression, which are recog-
nized initiators of involutio n. Th e sec ond ph ase was marked by
early adipoge nesis and prominent immune cell infiltratio n. Th e
ability of TIMP3 to differentially impact specific aspects of
inflammation that have been proposed during involution[26],
such as death receptors and ligands during the first phase and
cytokine/chemokines dependent immune cell influx during the
second phase, is revealed upon its combination with TNF
deficiency.
An early event in Timp3
2/2
mammary gland involution is E-
cadherin fragmentation and the loss of b-catenin resulting in the
loss of a critical survival signal that provides structural integrity.
Targeted deletion of E-cadherin in the mammary gland leads to
unscheduled apoptosis during pregnancy such that the gland
resembles involuting tissue [60]. Vallorasi et al. have shown that
Figure 6. Dependence of TNF on caspase 3 activation,
inflammatory and immune cell influx in involuting mammary
glands. Histomorphometric analyses of wild-type (WT; n = 4 mice),
Timp3
2/2
(T3
2/2
; n = 4), Tnf
2/2
(n = 4), and Timp3
2/2
Tnf
2/2
(n = 4)
involuting glands at days 1 and 3 of involution (1i and 3i): (A) cleaved
caspase 3; (B) F4/80-positive cells; and (C) CD3-positive cells. Bar graphs
are expressed as mean 6 SEM. * P,0.05; ** P,0.01.
doi:10.1371/journal.pone.0026718.g006
Figure 5. Increased and sustained inflammatory and immune
response in
Timp3
2
/
2
involuting glands. (A) Representative images
of F4/80-positive cells during wild-type (WT) and Timp3
2/2
involution at
day 10L and day 2i (B) Representative images of CD3
+
cells during
involution at 10L and 3i. (C) F4/80-postive and CD3-positive cell counts
as assessed by histomorphometry. (P values for F4/80: are 0.01 for 1i 2i
and 3i; P values for CD3: 1i, 0.04; 3i, 0.03; 5i, 0.02) for WT (n = 3) and
timp3
2/2
glands (n = 3). Bar graphs are expressed as mean 6 SEM. Scale
bar = 30 bm.
doi:10.1371/journal.pone.0026718.g005
TIMP3 Regulation of Mammary Involution
PLoS ONE | www.plosone.org 6 October 2011 | Volume 6 | Issue 10 | e26718
the truncation of E-cadherin is present at the initiation of
mammary involution [61]. Synthetic MMP inhibitors prevent E-
cadherin processing [50,62] and we have found increased MMP2
activation in a variety of Timp3 deficient systems [35,63,64]. In
particular, the immediate activation of MMP2 at the onset of
mammary involution in Timp3 deficient glands [23] coincides with
E-cadherin fragmentation, which is mitigated by reconstitution
with recombinant TIMP3. The presence of fragmentation
products at 10L suggests that epithelial structures may even be
compromised during lactation in Timp3 null mice. We propose
that the loss of E-cadherin triggers an impending change in
mammary architecture that sensitizes to apoptosis. Beyond
mammary gland involution, another in vivo model has demon-
strated that the loss of TIMP3 sensitizes tissues to cell death
pathways. During liver regeneration the loss of TIMP3 initiates an
apoptosis program due to excessive TNF [34]. In this study, TNF
dysregulation was concurrent with increased caspase 8 activation
and accompanied by an accelerated timeline of mitochondrial
apoptosis in Timp3
2/2
involuting mammary glands. The imme-
diate peak of apoptosis in Timp3
2/2
glands was attenuated by the
genetic ablation of TNF, thereby confirming the role of TNF in
the first phase of involution.
The loss of TIMP3 created an environment with greater
macrophage infiltration in comparison to wild-type mammary
glands. It is likely that this inflammatory influx does not initiate
apoptosis, because the peak of macrophage infiltration follows the
peak of apoptosis in both wild-type and Timp3
2/2
glands.
Apoptotic cell bodies must be removed by phagocytosis either by
neighboring epithelial cells or by professional phagocytes such as
macrophages in order to maintain tissue homeostasis during
mammary involution [65,66,67,68]. The number of CD3
+
T-cells
were also elevated and sustained in Timp3
2/2
glands following the
influx of macrophages. To date, lymphocyte infiltration has only
been described in bovine and ovine mammary gland involution
[56] and their function in the mammary gland remains to be
defined. We observed a pronounced macrophage and T-cell influx
upon the compound loss of TIMP3 and TNF, indicating a TNF-
independent mechanism for their recruitment into a TIMP3-
deficient microenvironment. Compound TIMP3 and TNF
deficiency in the heart also leads to a 5-fold greater neutrophil
infiltration in an otherwise non-inflammatory system [35]. Since
involution related inflammation has been theorized to play a role
in PABC, this separation of epithelial apoptosis from inflammation
provides insight into the mechanisms of inflammation during this
critical period.
Our results highlight the pleiotropic ability of the extracellular
metalloproteinase inhibitor TIMP3 during mammary gland
involution. TIMP3 influences distinct processes through TNF-
dependent and TNF-independent pathways with its loss acceler-
ating and amplifying specific involution events. TIMP3 acts as a
safeguard against untimely physiological cell death, unscheduled
inflammatory response, and premature loss of mammary gland
function. A greater understanding of apoptotic and inflammatory
regulators provides insight into PABC-related risk factors.
Materials and Methods
Ethics Statement
This study was carried out in strict accordance with the
recommendations in the Guide for the Care and Use of
Laboratory Animals of the National Institutes of Health with all
efforts made to minimize suffering. The protocol was approved by
the University Health Network Animal Care Committee (Animal
Use Protocol #812).
Figure 7. Schematic representation of the shifted timeline of involution in
Timp3
2
/
2
mammary glands. The timing and intensity of
epithelial apoptosis, adipogenesis, and immune cell influx are depicted by a solid line for wild-type and a dotted line for Timp3
2/2
bmammary gland
involution.
doi:10.1371/journal.pone.0026718.g007
TIMP3 Regulation of Mammary Involution
PLoS ONE | www.plosone.org 7 October 2011 | Volume 6 | Issue 10 | e26718
Mice and Experimental Involution
For the majority of the analyses, wild-type and Timp3
2/2
mice
were in FVB/N background. For the compound knockout studies
with TNF, wild-type, Timp3
2/2
, Tnf
2/2
, and Timp3
2/2
/Tnf
2/2
mice were all in C57/BL6 background. Forced experimental
involution was performed as previously described [23] according
to guidelines established by the Canadian Council for Animal
Care under protocols approved by the Ontario Cancer Institute
Animal Care Committee.
Antibodies and Reagents
The antibodies used were: Apaf-1 (Upstate); caspase 9 and
activated caspase 3 (Cell Signaling); E-cadherin, b-catenin,
cytochrome c, SMAC, and Mitotracker CD3, F4/80, B220 (BD
Biosciences); anti-neutrophil antibody (Serotec); AIF (gift from Dr.
J. Penninger); caspase 8 (gift from Dr. R. Hakem). Recombinant
human TIMP3 was prepared according to the manufacturer’s
specifications (R&D Systems).
Protein collection, western blotting, and TNF ELISA
Mammary gland collection and lysate preparation were
performed as previously described [23]. A total of 3 independent,
non-sibling females were used for each timepoint per genotype.
Protein quantification was performed using Dc Biorad Assay as
per manufacturer’s instructions. Equal loading was rigorously
assessed and adjusted for, based on amido black staining of newly
transferred gels and silver staining of SDS-PAGE gels that were
electrophoresed in parallel. The amido black-stained gels and
silver-stained gels were included and labeled as ‘Protein’ in the
Figures. For western blotting experiments, 50
mg of total protein
were loaded onto SDS-PAGE gels. Densitometry analyses were
performed using ImageQuant and Northern Eclipse software.
TNF ELISA was performed as previously described [34].
Pellet implantation
50 mg of recombinant human TIMP3 (R&D Systems) were
loaded onto pellets as per manufacturer’s instructions. Implanta-
tion of pellets was done as previously described [23].
Confocal microscopy and Immunocytochemistry
Two photon confocal microscopy was used for all immunoflu-
orescence and immunocytochemistry analyses. Exposure time,
gain, and offset parameters were first empirically determined and
then kept constant throughout. Preparation of mammary glands
for immunofluorescence and immunocytochemistry were per-
formed as previously described [23]. Depending on the primary
antibody used, control IgG antibodies and secondary antibodies
alone were used for controls to assess specificity of primary
antibodies.
Immunohistochemistry and Histomorphometry
Preparation of mammary glands for immunohistochemistry
were performed as previously described [23], using 1:50 dilutions
of anti-activated caspase 3 (Cell Signaling); anti-CD3, anti-F4/80,
and anti-B220 (BD Pharmingen); and anti-neutrophil (Serotec).
Histomorphometric counts of were performed on the fourth
inguinal gland, using 10 random fields distal to the lymph node.
Statistical Analyses
Statistical analyses were performed for observations that have at
least three mice per group. Student’s t test was performed for
statistical analyses between two groups (WT and Timp3
2/2
). One-
way ANOVA was performed for statistical comparison between
four groups (WT, Timp3
2/2
, Tnf
2/2
, and Timp3
2/2
/Tnf
2/2
),
followed by Tukey test for pair-wise statistical analysis.
Acknowledgments
We thank J.E. Fata for tissue collection at specific stages of mammary gland
involution and J.L. English and D.S. Smookler for ELISA analyses and P.
Waterhouse for critique of the manuscript.
Author Contributions
Conceived and designed the experiments: CVH RK. Performed the
experiments: CVH. Analyzed the data: CVH HWJ RK. Wrote the paper:
HWJ RK.
References
1. Garcia-Suarez O, Perez-Perez M, Germana A, Esteban I, Germana G (2003)
Involvement of growth factors in thymic involution. Microsc Res Tech 62:
514–523.
2. Kwong J, Choi HL, Huang Y, Chan FL (1999) Ultrastructural and biochemical
observations on the early changes in apoptotic epithelial cells of the rat prostate
induced by castration. Cell Tissue Res 298: 123–136.
3. Nilsen-Hamilton M, Liu Q, Ryon J, Bendickson L, Lepont P, et al. (2003) Tissue
involution and the acute phase response. Ann N Y Acad Sci 995: 94–108.
4. Howard BA, Gusterson BA (2000) Human breast development. J Mammary
Gland Biol Neoplasia 5: 119–137.
5. Green KA, Streuli CH (2004) Apoptosis regulation in the mammary gland. Cell
Mol Life Sci 61: 1867–1883.
6. Hennighausen L, Robinson GW (2005) Information networks in the mammary
gland. Nat Rev Mol Cell Biol 6: 715–725.
7. Wiseman BS, Werb Z (2002) Stromal effects on mammary gland development
and breast cance r. Science 296: 1046–1049.
8. Alexander CM, Selvarajan S, Mudgett J, Werb Z (2001) Stromelysin-1 regulates
adipogenesis during mammary gland involution. J Cell Biol 152: 693–703.
9. Clarkson RW, Wayland MT, Lee J, Freeman T, Watson CJ (2004) Gene
expression profiling of mammary gland development reveals putative roles for
death receptors and immune mediators in post-lactat ional regression. Breast
Cancer Res 6: R92–R109.
10. Stein T, Morris JS, Davies CR, Weber-Hall SJ, Duffy MA, et al. (2004)
Involution of the mouse mammary gland is associated with an immune cascade
and an acute-phase response, involving LBP, CD14 and STAT3. Breast Cancer
Res 6: R75–R91.
11. Lund LR, Romer J, Thomasset N, Solberg H, Pyke C, et al. (1996) Two distinct
phases of apoptosis in mammary gland involution: proteinase-independent and -
dependent pathways. Development 122: 181–193.
12. Stein T, Salomonis N, Gusterson BA (2007) Mammary gland involution as a
multi-step process. J Mammary Gland Biol Neoplasia 12: 25–35.
13. Varela LM, Ip MM (1996) Tumor necrosis factor-alpha: a multifunctional
regulator of mammary gland development. Endocrinology 137: 4915–4924.
14. Song J, Sapi E, Brown W, Nilsen J, Tartaro K, et al. (2000) Roles of Fas and Fas
ligand during mammary gland remodeling. J Clin Invest 106: 1209–1220.
15. Nguyen AV, Pollard JW (2000) Transforming growth factor beta3 induces cell
death during the first stage of mammary gland involution. Development 127:
3107–3118.
16. Humphreys RC, Bierie B, Zhao L, Raz R, Levy D, et al. (200 2) Deletion of Stat3
blocks mammary gland involution and extends functional competence of the
secretory epithelium in the absence of lactogenic stimuli. Endocrinology 143:
3641–3650.
17. Chapman RS, Lourenco PC, Tonner E, Flint DJ, Selbert S, et al. (1999)
Suppression of epithelial apoptosis and delayed mammary gland involution in
mice with a conditional knockout of Stat3. Genes Dev 13: 2604–2616.
18. Marti A, Ritter PM, Jager R, Lazar H, Baltzer A, et al. (2001) Mouse mammary
gland involution is associated with cytochrome c release and caspase activation.
Mech Dev 104: 89–98.
19. Fata JE, Werb Z, Bissell MJ (2004) Regulation of mammary gland branching
morphogenesis by the extracellular matrix and its remodeling enzymes. Breast
Cancer Res 6: 1–11.
20. Fata JE, Leco KJ, Moorehead RA, Martin DC, Khokha R (1999) Timp-1 is
important for epithelial proliferation and branching morphogenesis during
mouse mammary development. Dev Biol 211: 238–254.
21. Wiseman BS, Sternlicht MD, Lund LR, Alexander CM, Mott J, et al. (2003)
Site-specific inductive and inhibitory activities of MMP-2 and MMP-3
orchestrate mammary gland branching morphogenesis. J Cell Biol 162:
1123–1133.
TIMP3 Regulation of Mammary Involution
PLoS ONE | www.plosone.org 8 October 2011 | Volume 6 | Issue 10 | e26718
22. Sympson CJ , Talhouk RS, Alexander CM, Chin JR, Clift SM, et al. (1994)
Targeted expression of stromelysin-1 in mammary gland provides evidence for a
role of proteinases in branching morphogenesis and the requirement for an
intact basement membrane for tissue-specific gene expression. J Cell Biol 125:
681–693.
23. Fata JE, Leco KJ, Voura EB, Yu HY, Waterhouse P, et al. (2001) Accelerated
apoptosis in the Timp-3-deficient mammary gland. J Clin Invest 108: 831–841.
24. Khokha R, Werb Z (2011) Mammary gland reprogramming: metalloproteinases
couple form with function. Cold Spring Harb Perspect Biol 3.
25. Talhouk RS, Bissell MJ, Werb Z (1992) Coordinated expression of extracellular
matrix-degrading proteinases and their inhibitors regulates mammary epithelial
function during involution. J Cell Biol 118: 1271–1282.
26. Watson CJ (2009) Immune cell regulators in mouse mammary development and
involution. J Anim Sci 87: 35–42.
27. O’Brie n J, Lyons T, Monks J, Lucia MS, Wilson RS, et al. (2010) Alternatively
activated macrophages and collagen remodeling characterize the postpartum
involuting mammary gland across species. Am J Pathol 176: 1241–1255.
28. Schedin P (2006) Pregnancy-associated breast cancer and metastasis. Nat Rev
Cancer 6: 281–291.
29. Lin EY, Nguyen AV, Russell RG, Pollard JW (2001) Colony-stimulating factor 1
promotes progression of mammary tumors to malignancy. J Exp Med 193:
727–740.
30. Lyons TR, Schedin PJ, Borges VF (2009) Pregnancy and breast cancer: when
they collide. J Mammary Gland Biol Neoplasia 14: 87–98.
31. Stensheim H, Moller B, van Dijk T, Fossa SD (2009) Cause-specific survival for
women diagnosed with cancer during pregnancy or lactation: a registry-based
cohort study. J Clin Oncol 27: 45–51.
32. Math elin C, Annane K, Treisser A, Chenard MP, Tomasetto C, et al. (2008)
Pregnancy and post-partum breast cancer: a prospective study. Anticancer Res
28: 2447–2452.
33. Woessner JF, Jr. (2001) That impish TIMP: the tissue inhibitor of
metalloproteinases-3. J Clin Invest 108: 799–800.
34. Mohamm ed FF, Smookler DS, Taylor SE, Fingleton B, Kassiri Z, et al. (2004)
Abnormal TNF activity in Timp3
-/-
mice leads to chronic hepatic inflammation
and failure of liver regeneration. Nat Genet 36: 969–977.
35. Kassiri Z, Oudit GY, Sanchez O, Dawood F, Mohammed FF, et al. (2005)
Combination of tumor necrosis factor-alpha ablation and matrix metalloprotei-
nase inhibition prevents heart failure after pressure overload in tissue inhibitor of
metalloproteinase-3 knock-out mice. Circ Res 97: 380–390.
36. Mahmoodi M, Sahebjam S, Smookler D, Khokha R, Mort JS (2005) Lack of
tissue inhibitor of metalloproteinases-3 results in an enhanced inflammatory
response in antigen-induced arthritis. Am J Pathol 166: 1733–1740.
37. Smookle r DS, Mohammed FF, Kassiri Z, Duncan GS, Mak TW, et al. (2006)
Tissue inhibitor of metalloproteinase 3 regulates TNF-dependent systemic
inflammation. J Immunol 176: 721–725.
38. Amour A, Slocombe PM, Webster A, Butler M, Knight CG, et al. (1998) TNF-
alpha converting enzyme (TACE) is inhibited by TIMP-3. FEBS Lett 435:
39–44.
39. Lee MH, Knauper V, Becherer JD, Murphy G (2001) Full-length and N-TIMP-
3 display equal inhibitory activities toward TNF-alpha convertase. Biochem
Biophys Res Commun 280: 945–950.
40. Varfolo meev EE, Schuchmann M, Luria V, Chiannilkulchai N, Beckmann JS,
et al. (1998) Targeted disruption of the mouse Caspase 8 gene ablates cell death
induction by the TNF receptors, Fas/Apo1, and DR3 and is lethal prenatally.
Immunity 9: 267–276.
41. Zou H, Henzel WJ, Liu X, Lutschg A, Wang X (1997) Apaf-1, a human protein
homologous to C. elegans CED-4, participates in cytochrome c-dependent
activation of caspase-3. Cell 90: 405–413.
42. Li P, Nijhawan D, Budihardjo I, Srinivasula SM, Ahmad M, et al. (1997)
Cytochrome c and dATP-dependent formation of Apaf-1/caspase-9 complex
initiates an apoptotic protease cascade. Cell 91: 479–489.
43. Liu X, Kim CN, Yang J, Jemmerson R, Wang X (1996) Induction of apoptotic
program in cell-free extracts: requirement for dATP and cytochrome c. Cell 86:
147–157.
44. Green DR, Kroemer G (2004) The pathophysiology of mitochondrial cell death.
Science 305: 626–629.
45. Okada H, Suh WK, Jin J, Woo M, Du C, et al. (2002) Generation and
characterization of Smac/DIABLO-deficient mice. Mol Cell Biol 22:
3509–3517.
46. Zhao L, Melenhorst JJ, Hennighausen L (2002) Loss of interleukin 6 results in
delayed mammary gland involution: a possible role for mitogen-activated
protein kinase and not signal transducer and activator of transcription 3. Mol
Endocrinol 16: 2902–2912 .
47. Schere-Levy C, Buggiano V, Quaglino A, Gattelli A, Cirio MC, et al. (2003)
Leukemia inhibitory factor induces apoptosis of the mammary epitheli al cells
and participates in mouse mammary gland involution. Exp Cell Res 282: 35–47.
48. Yeh WC, Cao Z, Classon M, McKnight SL (1995) Cascade regulation of
terminal adipocyte differentiation by three members of the C/EBP family of
leucine zipper proteins. Genes Dev 9: 168–181.
49. Ross SE, Hemati N, Longo KA, Bennett CN, Lucas PC, et al. (2000) Inhibition
of adipogenesis by Wnt signaling. Science 289: 950–953.
50. Noe V, Fingleton B, Jacobs K, Crawford HC, Vermeulen S, et al. (2001) Release
of an invasion promoter E-cadherin fragment by matrilysin and stromelysin-1.
J Cell Sci 114: 111–118.
51. Lochter A, Galosy S, Muschler J, Freedman N, Werb Z, et al. (1997) Matrix
metalloproteinase stromelysin-1 triggers a cascade of molecular alterations that
leads to stable epithelial-to-mesenchymal conversion and a premalignant
phenotype in mammary epithelial cells. J Cell Biol 139: 1861–1872.
52. Monks J, Geske FJ, Lehman L, Fadok VA (2002) Do inflammatory cells
participate in mammary gland involution? J Mammary Gland Biol Neoplasia 7:
163–176.
53. Lee CS, McDowell GH, Lascelles AK (1969) The importance of macrophages in
the removal of fat from the involuting mammary gland. Res Vet Sci 10: 34–38.
54. Hurley WL, Finkelstein E (1986) Bovine leukocytes in mammary secretions
during involution: surface protein changes. Am J Vet Res 47: 2418–2422.
55. Black RA (2004) TIMP3 ch ecks inflammation. Nat Genet 36: 934–935.
56. Tatarczuch L, Philip C, Bischof R, Lee CS (2000) Leucocyte phenotypes in
involuting and fully involuted mammary glandular tissues and secretions of
sheep. J Anat 196(Pt 3): 313–326.
57. Wilson RA, Linn JA, Eberhart RJ (1986) A study of bovine T-cell subsets in the
blood and mammary gland secretions during the dry period. Vet Immunol
Immunopathol 13: 151–164.
58. Murthy A, Defamie V, Smookler DS, Di Grappa MA, Horiuchi K, et al. (2010)
Ectodomain shedding of EGFR ligands and TNFR1 dictates hepatocyte
apoptosis during fulminant hepatitis in mice. J Clin Invest 120: 2731–2744.
59. Watson CJ (2006) Key stages in mammary gland development - Involution:
apoptosis and tissue remodelling that convert the mammary gland from milk
factory to a quiescent organ. Breast Cancer Res 8: 203.
60. Boussadia O, Kutsch S, Hierholzer A, Delmas V, Kemler R (2002) E-cadherin is
a survival factor for the lactating mouse mammary gland. Mech Dev 115: 53–62.
61. Vallorosi CJ, Day KC, Zhao X, Rashid MG, Rubin MA, et al. (2000)
Truncation of the beta-catenin binding domain of E-cadherin precedes epithelial
apoptosis during prostate and mammary invo lution. J Biol Chem 275:
3328–3334.
62. Steinhusen U, Weiske J, Badock V, Tauber R, Bommert K, et al. (2001)
Cleavage and shedding of E-cadherin after induction o f apoptosis. J Biol Chem
276: 4972–4980.
63. Cruz-Munoz W, Sanchez OH, Di Grappa M, English JL, Hill RP, et al. (2006)
Enhanced metastatic disseminat ion to multiple organs by melanoma and
lymphoma cells in timp-3(
-/-
) mice. Oncogene.
64. English JL, Kassiri Z, Koskivirta I, Atkinson SJ, Di Grappa M, et al. (2006)
Individual timp deficiencies differentially impact pro-MMP-2 activaiton. J Biol
Chem.
65. Atabai K, Fernandez R, Huang X, Ueki I, Kline A, et al. (2005) Mfge8 is critical
for mamma ry gland rem o del in g d uri ng i n vo lu tio n. M ol B io l Cel l 16 :
5528–5537.
66. Sandahl M, Hunter DM, Strunk KE, Earp HS, Cook RS (2010) Epithelial cell-
directed efferocytosis in the post-partum mammary gland is necessary for tissue
homeostasis and future lactation. BMC Dev Biol 10: 122.
67. Monks J, Rosner D, Geske FJ, Lehman L, Hanson L, et al. (2005) Epithelial cells
as phagocytes: apoptotic epithelial cells are engulfed by mammary alveolar
epithelial cells and repress inflammatory mediator release. Cell Death Differ 12:
107–114.
68. Hanayama R, N agata S (2005) Impaired involution of mammary glands in the
absence of milk fat globule EGF factor 8. Proc Natl Acad Sci U S A 102:
16886–16891.
TIMP3 Regulation of Mammary Involution
PLoS ONE | www.plosone.org 9 October 2011 | Volume 6 | Issue 10 | e26718
