Download Cardiac troponin I gene knockout - University of Wisconsin–Madison

Original Contributions
Cardiac Troponin I Gene Knockout
A Mouse Model of Myocardial Troponin I Deficiency
XuPei Huang, YeQing Pi, Kevin J. Lee, Anne S. Henkel, Ronald G. Gregg,
Patricia A. Powers, Jeffery W. Walker
Abstract—Troponin I is a subunit of the thin filament–associated troponin-tropomyosin complex involved in calcium
regulation of skeletal and cardiac muscle contraction. We deleted the cardiac isoform of troponin I by using gene
targeting in murine embryonic stem cells to determine the developmental and physiological effects of the absence of this
regulatory protein. Mice lacking cardiac troponin I were born healthy, with normal heart and body weight, because a
fetal troponin I isoform (identical to slow skeletal troponin I) compensated for the absence of cardiac troponin I.
Compensation was only temporary, however, as 15 days after birth slow skeletal troponin I expression began a steady
decline, giving rise to a troponin I deficiency. Mice died of acute heart failure on day 18, demonstrating that some form
of troponin I is required for normal cardiac function and survival. Ventricular myocytes isolated from these troponin
I– depleted hearts displayed shortened sarcomeres and elevated resting tension measured under relaxing conditions and
had a reduced myofilament Ca sensitivity under activating conditions. The results show that (1) developmental
downregulation of slow skeletal troponin I occurs even in the absence of cardiac troponin I and (2) the resultant troponin
I depletion alters specific mechanical properties of myocardium and can lead to a lethal form of acute heart failure. (Circ
Res. 1999;84:1-8.)
Key Words: ischemia n heart failure n cardiac development
C
ontraction of heart muscle is triggered by calcium
binding to the troponin-tropomyosin regulatory complex
on cardiac muscle filaments. One protein subunit of this
regulatory complex is troponin I (TnI), a 24-kDa phosphoprotein thought to play a central role in communicating
calcium binding to activation of the actin filament.1–3 Understanding TnI function in the heart is important, because TnI
may be modified in certain cardiac-diseased states. Selective
proteolysis of TnI has been reported to underlie the pathology
of stunned myocardium.4 Moreover, TnI may become depleted in ischemic,5 infarcted,6 – 8 and possibly failing myocardium.9 A cardiac-specific TnI isoform (cTnI) is released
from damaged myocardial tissue, providing a clinically useful
serum marker for diagnosis of some acute coronary (or
ischemic) syndromes.10 Importantly, this TnI loss may not be
restricted to necrotic myocardial tissue but may also occur in
tissue that survives and attempts to maintain organ function.7–9 Interpretation of the effects of TnI loss in animal
models of cardiac disease is often complicated by other
cellular changes that occur in parallel.5– 8 On the other hand,
selective removal of TnI has been reported,11,12 but this
approach limits investigation to permeabilized fiber bundles.
A more complete understanding of the physiological consequences of myocardial TnI depletion in vivo is needed.
To better understand and control TnI expression in the
heart, we knocked out the gene for cTnI in mice. Two main
TnI genes are expressed in the mammalian heart under the
control of a developmentally regulated program.13 Fetal TnI,
which is identical to slow skeletal TnI (ssTnI), is expressed
first and predominates throughout embryonic and fetal development. Around embryonic day 10, cTnI begins to be
expressed. Soon after birth, cTnI accounts for roughly half of
the total thin-filament TnI content, and then cTnI predominates throughout adulthood in most mammals, including
humans.13–15 Since the mechanism of the fetal to adult TnI
isoform switch is not well understood, we considered 2
possible outcomes of deleting the cTnI gene. The fetal ssTnI
isoform could compensate, resulting in effective replacement
of cTnI with ssTnI in the adult mouse heart. Alternatively,
ssTnI could be downregulated as usual during the TnI
isoform switch, creating a mammalian model with a myocardial TnI deficiency.
Homologous recombination in embryonic stem (ES) cells
was used to delete the entire cTnI gene from the mouse
genome. ssTnI was found to compensate for the absence of
cTnI, but only temporarily, as ssTnI was eventually downregulated by an as-yet-unidentified mechanism. This provided an opportunity to assess the effects on cardiac function
of TnI loss in a system that should be relatively free of other
Received July 14, 1998; accepted October 22, 1998.
From the Department of Physiology, University of Wisconsin, Madison, Wis. The current affiliation for Dr Gregg is the Departments of Biochemistry
and Ophthalmology, University of Louisville, Louisville, Ky.
Correspondence to Dr Jeffery W. Walker, Department of Physiology, 1300 University Ave, Madison, WI 53706. E-mail [email protected]
© 1999 American Heart Association, Inc.
Circulation Research is available at http://www.circresaha.org
1
2
Cardiac Troponin I Knockout
confounding cellular changes known to occur in cardiacdiseased states, such as altered TnI phosphorylation8,16 and
defective excitation-contraction coupling. 17 The results
clearly show that TnI depletion alters both resting and active
mechanical properties of ventricular myocytes and causes a
lethal phenotype when TnI content drops below a critical
threshold. The results also show for the first time that
expression of the adult cTnI isoform is not an essential
component of the switch that terminates expression of the
fetal ssTnI isoform in the developing heart. This mouse strain
with its predictable time-dependent loss of TnI should be
useful in future investigations to establish the effects of TnI
depletion on cardiac function in vivo.
Materials and Methods
Gene Targeting
A P1 clone containing the mouse cardiac TnI gene was obtained
from a mouse 129 P1 genomic library (Genome Systems, St. Louis,
MO). An 11-kb SalI fragment from the P1 clone was isolated,
subcloned, and found to contain the entire cardiac TnI gene (Figure
1A). A 2.5-kb SalI/BamHI fragment located upstream of exon 1 and
a 2.2-kb BamHI fragment located downstream of exon 8 were used
as the 59 and 39 homology units, respectively. These fragments were
cloned into a pNT3 vector containing a neomycin (neo) resistance
gene linked to a thymidine kinase promoter flanked by multiple
cloning sites. The vector also contained a herpes simplex thymidine
kinase cassette for negative selection of nonrecombinant ES cell
clones. The resulting targeting vector (Figure 1A) contained a unique
SalI restriction site for linearization of the plasmid before electroporation into ES cells. Homologous recombination between the
targeting vector and the cognate cTnI locus deleted exons 1 to 8, or
the entire cTnI allele (Figure 1A). Colony selection and target clone
identification were as described.18 Targeting vector (25 mg) was
introduced into 53106 ES cells. Of 256 G418 and 1-(29-deoxy-29fluoro-b-D-arabinofuranosyl)-5-iodouracil–resistant clones, 96 were
analyzed and 12 were found to be correctly targeted by Southern
blotting with probe 1 and subsequently with probe 2. Three correctly
targeted clones were microinjected into C57B1/6J blastocysts, and
blastocysts were implanted in pseudopregnant 129 mice. Appropriate
breeding was used to produce mice either heterozygous or homozygous for the targeted cTnI allele. Southern blot analysis was initially
used for genotyping (Figure 1B). Genomic DNA was isolated from
tail biopsies using Purgene DNA isolation (Gentra Systems). Five
micrograms of genomic DNA was digested and then size fractionated on agarose gels, transferred to nylon membranes, and probed
with DNA labeled with 32P by a random oligonucleotide protocol.
Probe 1 detected a 6.5-kb EcoRI fragment from the wild-type cTnI
allele and a 4.8-kb EcoRI fragment from the targeted allele (Figure
1B). A polymerase chain reaction (PCR)– based assay was also
developed for rapid screening. Primers were designed to produce a
630-bp and a 390-bp fragment for the wild-type and targeted alleles,
respectively.
loads were standardized by a bicinchoninic acid protein assay before
electrophoresis and by quantitative densitometry of Coomassie
blue–stained gels as described.19
Histology
Care and handling of mice were carried out according to institutional
guidelines approved by the Association for Assessment and Accreditation of Laboratory Care (AAALAC) International. Mice were
killed by cervical dislocation. For light microscopy, hearts were
quickly excised and immersed in 10% formaldehyde solution at
room temperature. Fixed hearts were sectioned into 50-mm-thick
slices, stained with hematoxylin and eosin, and then viewed under a
Nikon Diaphot inverted microscope equipped with a 53 objective
and National Institutes of Health Image software. Lung tissue was
rapidly frozen in liquid nitrogen after lungs were inflated with air to
'50 mm hydrostatic pressure. Frozen tissue was sectioned into 5- to
10-mm-thick slices and viewed under the light microscope. For
electron microscopy (EM) analysis, excised hearts were rapidly
immersed in PBS containing 2% (vol/vol) paraformaldehyde and 2%
(vol/vol) glutaraldehyde. Blocks of tissue 1 mm2 were dissected,
embedded in resin, sectioned, and viewed on a Philips CM120
transmission electron microscope.
Force Measurements
Excised hearts were depleted of blood by massage in Ringer’s
solution, and then ventricular tissue was diced and homogenized for
3 to 5 s in 5 mL of relaxing solution using a Polytron homogenizer.
Cells were collected by differential centrifugation on a tabletop
centrifuge and then incubated for 6 minutes at room temperature in
relaxing solution containing 0.3% Triton X-100 and 0.5 mg/mL
BSA. Skinned single myocytes (or small bundles of up to 4 cells)
were attached to a Cambridge model 403 force transducer, and force
output was monitored as described.20 A passive length-tension
relationship was determined in relaxing solution containing no added
calcium (nominally pCa 9, where pCa-log[Ca21]). Sarcomere length
(SL) was monitored by video microscopy and varied from 1.7 to
2.7 mm in 0.2 to 0.4 –mm increments. Active tension was recorded at
2.2 mm by transferring the attached myocyte into activating solution
at pCa 4.5 (32 mmol/L free Ca21) as described.20 All tension
measurements were carried out at 20°C to 22°C and normalized to
myocyte cross-sectional area.
Solutions
The composition of Ringer’s solution was as follows (in mmol/L):
NaCl 118, KCl 4.8, HEPES 25 (adjusted to pH 7.4 with NaOH),
KH2PO4 2, MgCl2 1.2, pyruvate 5, and glucose 11. The composition
of relaxing solution was (in mmol/L): KCl 80, MgATP 4, MgCl2 7.4
(1 free Mg21), EGTA 7, creatine phosphate 11, and imidazole 25
(adjusted to pH 7.0 with KOH). Measurements of maximum active
tension were performed with an activating solution of essentially the
same composition as relaxing solution but containing CaCl2 (added
before pH adjustment) to achieve 32 mmol/L free Ca (pCa 4.5), and
KCl was adjusted to maintain a constant ionic strength of 0.18 mol/L.
For intermediate Ca levels between pCa 4.5 and 9, a mixing table
was used to determine the amount of relaxing and activating solution
to achieve pCa values of 4.8, 5.0, 5.2, 5.5, 5.8, 6.0, and 6.5.20
Immunoblotting
TnI protein levels were determined by Western blotting of ventricular tissue homogenates after electrophoresis in 12% SDS-PAGE
gels and transfer onto nitrocellulose paper. An anti-TnI monoclonal
antibody (clone 6F9, Advanced ImmunoChemical Inc) that recognized both mouse cTnI and ssTnI was used at a dilution of 1:1000.
The identities of positive bands were confirmed by isoform-specific
TnI monoclonal antibodies from the same commercial source (clone
7F4 for cTnI and clone A9 for ssTnI). ssTnI was further confirmed
by comigration with an antibody-positive band (clones A9 and 6F9)
in mouse soleus muscle. Antibodies on immunoblots were visualized
by enhanced chemiluminescence (ECL). For quantification of TnI
content, ECL bands were scanned by densitometry and compared
with a range of pure cTnI on the same blot to ensure linearity. Protein
Statistics
Data are expressed as mean6SEM and statistically analyzed by both
paired and unpaired Student’s t tests. P,0.05 was considered to be
a significant difference. Data fitting was carried out by use of
Marquardt’s nonlinear regression method. The quality of fits was
taken to be acceptable when the SEs in the estimated pCa50 and nH
values were #12% and P,0.005.
Results
The cTnI gene was deleted by using homologous recombination in murine ES cells. A vector containing a neomycin resistance cassette was used for positive selection of
Huang et al
January 8/22, 1999
Figure 1. Top, Structure and detection of cTnI before and after gene targeting. A, cTnI gene structure, restriction sites for EcoRI (RI)
and SalI (S), and location of probes for Southern blots. Middle, Targeting vector with positive (NEO) and negative (thymidine kinase
[TK]) selectable markers. Bottom, cTnI knockout and expected PCR and Southern blot fragments. B, Top, Southern blot using DNA
probe 1 of mouse genomic DNA obtained from tail biopsies of a litter of 14-day-old pups. Genotypes are indicated under each animal
number. Bottom, Western blot of ventricular samples from the same pups using a monoclonal antibody that recognizes both cTnI and
ssTnI.
3
4
Cardiac Troponin I Knockout
TABLE 1.
Summary of Mouse Phenotypes
Weight*
Mortality, %
Body, g
Heart, mg
Heart/Body, %
Day 17
Day 18
Homozygous (cTnI2/2)
8.861.4
47.462.4
0.5460.03
7.8
92.2
Heterozygous (cTnI1/2)
8.162.3
42.369.7
0.5160.12
0
0
Wild type
8.461.3
44.766.8
0.5360.08
0
0
*Data are from mice on day 18 after birth and are reported as mean6SEM with n$20.
targeted ES cells, and a thymidine kinase gene was used
for negative selection against random insertion (Figure 1).
Mice heterozygous for targeted cTnI (cTnI1/–) were
indistinguishable from wild-type mice in body size, general appearance, and life span (Table 1). Heterozygote
crosses showed normal fertility (60/64 pairs, 94%) with
low neonatal mortality (8/448, 1.8%) and produced normal
litter sizes (5 to 14 pups/litter, mean57.462.6). The
genotype distribution of live neonates was 27:48:25 (wild
type:cTnI1/–:cTnI–/–) as determined by Southern blots
(Figure 2, top), a ratio consistent with mendelian segregation. The absence of the cTnI protein in cTnI–/– animals
was confirmed by immunoblots with anti-TnI antibodies
(Figure 2, bottom). Remarkably, cTnI–/– animals were
also indistinguishable from wild-type animals on the basis
of body weight and heart/body weight ratio measured up to
18 days postnatally (Table 1).
A time course of protein expression levels showed the
expected fetal to adult TnI isoform switch in wild-type mouse
hearts, but the switch was delayed by '3 days in heterozygotes (cTnI1/–) (Figure 2). In cTnI–/– mice, ssTnI remained
elevated for at least 10 days beyond the normal time taken for
the isoform switch (Figure 2), apparently compensating for
the absence of adult cTnI and accounting for the normal
appearance of cTnI–/– animals. At day 15 after birth, ssTnI
levels began an abrupt decline that continued over the next 3
days, reaching 3669% of its original value (Figure 2). During
this decline, there was no evidence of proteolytic fragments
of TnI with the antibodies used. On day 18, dyspnea (difficulty breathing) and lethargy developed, and cTnI–/– animals
died within hours of the onset of these overt signs. Both the
extreme nature and the consistency of the phenotype were
striking, as .90% of cTnI–/– animals died on day 18 after
birth, with the other 10% having died on day 17 (Table 1).
Initiation of a hypertrophic growth program and other
forms of remodeling including infiltration and fibrosis are
common responses to cardiac insufficiency.21,22 Heart tissue
sections prepared from cTnI–/– animals and stained with
hematoxylin and eosin revealed neither hypertrophy nor
ventricular dilation even in 18-day-old animals with advanced symptoms (not shown). Gomori’s trichrome staining
for fibrosis also showed no differences between wild-type
and cTnI-null hearts (not shown). Examination of ventricular
tissue from 17- to 18-day-old cTnI–/– mice by EM showed
few changes in muscle ultrastructure. Myofibrillar disarray,
fibrosis, and macrophage infiltration were minimal. The lack
of obvious indications of remodeling in cTnI–/– mice may be
a characteristic of this type of myofilament defect, or more
likely it is a reflection of how suddenly TnI loss occurs.
Regardless of the correct explanation, these observations
suggest that TnI depletion is the principal defect in cTnI–/–
hearts and that secondary effects, because of tissue remodeling, contribute little to the pathology. Comparison of wildtype and null cardiac myofibrils by SDS-PAGE with silver
Figure 2. Time course of the TnI isoform
switch. Anti-TnI immunoblots of cardiac
muscle homogenates obtained from
wild-type (left panels), heterozygote
(cTnI1/–) (middle panels), and cTnI–/–
mice (right panels) euthanized at different
times after birth. Each lane contains 20
mg of ventricular protein. Graphs show
time courses of mean isoform expression levels taken from 3 separate experiments. The transition point is the estimated day on which ssTnI represented
half of the total TnI.
Huang et al
January 8/22, 1999
5
Figure 3. EM of ventricular tissue. Images showing ultrastructure of ventricular tissue from wild-type (top) or 17-day-old
cTnI–/– (bottom) animals. Arrows indicate adjacent Z-lines.
staining revealed no obvious differences in protein profiles
(other than TnI), but given uncertainties in this analysis, we
cannot rule out the possibility that other myofilament proteins
were altered as a consequence of TnI depletion.
Analysis of myofibrillar SL in EM sections did reveal 1
significant change in cTnI–/– animals. SL measured from
Z-line to Z-line was greatly reduced in cTnI–/– mice compared with wild-type littermates (cTnI–/– SL51.260.2 mm,
n510; wild-type SL52.160.1 mm, n510) (Figure 3). Shortening of sarcomeres was not observed in cTnI–/– mice up to
day 14 after birth, but developed later, in parallel with the loss
of ssTnI. This shortening of sarcomeres under relaxed conditions indicates the presence of Ca-independent forces in the
direction of normal active muscle contraction. The nature of
this force is currently unclear, but EM images showed
changes in mitochondria including a 4769% (P,0.01) increase in numbers and a trend toward larger individual
mitochondria. These observations are consistent with an
increase in O2/ATP consumption, suggesting that ATPdependent active forces are responsible for shortened sarcomeres in Figure 3.
Direct functional measurements in isolated hearts have
thus far been precluded by the small size of 17- to 18-day-old
mouse hearts. Histological analysis of lung tissue revealed
grossly enlarged and congested pulmonary capillaries (not
shown), consistent with left ventricular failure. To determine
the functional defects at the cellular level, cardiac myocytes
were isolated and attached to a force transducer after the
surface membrane was removed by detergent skinning.20 In
Figure 4. Mechanical properties of isolated ventricular myocytes. A, Isometric force under relaxing conditions as a function
of SL in cTnI–/– myocytes at day 14 ( f ) and day 17 (Œ) after
birth. F, Data from 17-day-old wild-type myocytes. Data are
mean6SEM. #P,0.02; *P,0.005. Solid lines are fits to
f(x)5A*exp[B*(x–xo)/xo)] (Reference 34), with A50.79, B52.92,
and xo51.52 mm (Œ), and A50.28, B53.45, and xo51.58 mm ( f
and F). B, Ca sensitivity of active tension development. Normalized tension-pCa relationships are shown for cTnI–/– myocytes
at day 14 ( f ) and day 17 (Œ) after birth. Solid lines represent
fits to the Hill equation: f(x)51/{112.3*exp[nH*(pCa–pCa50)]}.
myocytes from 17-day-old cTnI–/– mice, isometric tension
measured under relaxing conditions was significantly elevated compared with wild-type myocytes at all SLs examined
(Figure 4A). Resting tension was not elevated in 14-day-old
cTnI–/– myocytes (Figure 4A), so the increased resting
tension at 17 days cannot be attributed to the presence of
ssTnI rather than cTnI, but it is clearly the result of TnI
depletion. These steady-state force measurements directly
demonstrate the presence of a force responsible for the
shortened sarcomeres observed in ventricular tissue sections.
Finally, the effects of TnI depletion on Ca-activated
tension were examined. To assess the effects of TnI loss,
tension-pCa curves were compared in myocytes from 14-dayold and 17-day-old cTnI–/– mice (Figure 4B). During this
3-day period, TnI content was reduced from its maximum
level to ,40% of maximum. This loss of TnI did not greatly
6
Cardiac Troponin I Knockout
TABLE 2.
Summary of Active Tension in Skinned Myocytes
Age, d
TnI Content*
Po,† mN/mm2
pCa50‡
nH‡
cTnI2/2
14
9368% ssTnI
7.460.8
5.7060.03
1.5660.20
cTnI2/2
17
3669% ssTnI
9.860.5
5.3660.03
1.4060.14
Wild type
14
.90% cTnI
8.260.9
5.1960.04
1.5660.17
Wild type
17
.90% cTnI
12.860.9
5.1060.05
1.7860.20
*For cTnI2/2 animals, ssTnI content is expressed as percentage of total TnI in wild-type
age-matched littermates (mean6SEM, n55); cTnI content is 0%. cTnI content in wild-type animals
was not quantified with precision because of the likelihood that very low levels of ssTnI and high
levels of cTnI were not in the linear range for detection by ECL with a single exposure.
†Total tension (active plus resting) measured at pCa 4.5, 22°C. Data are reported as mean6SEM
(n58) for all tension measurements.
‡Obtained from fits to the Hill equation of tension-pCa data, as in Figure 4B.
affect maximum tension (Po) because, although Po increased
somewhat, a similar increase was observed in wild-type
myocytes over this same time period (Table 2). The largest
effect of TnI loss was on the pCa50 value, which decreased
from pCa 5.70 to 5.36 (Figure 4B). It is not likely that
developmental changes in other myofilament proteins contributed to this time-dependent change in pCa50, because
shifts in the tension-pCa relationship were considerably less
in wild-type myocytes (Table 2). Thus, TnI depletion is
associated with a decrease in the Ca sensitivity of tension
development in this system.
Discussion
We deleted the cardiac-specific TnI gene, cTnI, with the goal
of manipulating the content or the species of TnI expressed in
the mouse heart. Because there are 2 main TnI genes
expressed in the mammalian heart, ssTnI and cTnI (located
on different chromosomes), it was unclear how ssTnI expression would change in response to deletion of cTnI. In
wild-type ventricular tissue, ssTnI and cTnI are reciprocally
expressed, giving rise to a time-dependent isoform switch, as
follows. ssTnI is expressed first during development and then
is gradually replaced by cTnI, beginning at about embryonic
day 10 and finishing around postnatal day 10.13–15 At one
extreme, the ssTnI expression pattern could have been completely unaffected by the absence of cTnI and the heart would
be effectively depleted of ssTnI by postnatal day 10. At the
other extreme, ssTnI could have remained at a high level of
expression throughout development and growth, thereby fully
compensating for the absence of cTnI. What we observed was
somewhere between these extremes. ssTnI initially compensated for the absence of cTnI, but beginning around day 15
after birth a steady loss of ssTnI occurred, giving rise to a
condition of TnI deficiency.
It is clear from the present results that TnI depletion has
deleterious effects on the mechanical properties of cardiac
muscle, including changes in both resting and active tension.
One observed defect was elevation of resting tension, but the
nature of this force in the direction of sarcomere shortening is
presently unclear. TnI extraction experiments in isolated
ventricular trabeculae have shown that active cross-bridges
are recruited (even in the absence of Ca) as TnI is removed.11
In the present study, EM sections of cTnI-null ventricular
tissue revealed proliferation and enlargement of mitochon-
dria, suggesting increased ATP consumption, consistent with
the possibility that the force is due to recruitment of unregulated active cross-bridges as TnI is lost. In this respect the
effect of TnI depletion appears to be similar to cTnI extraction in skinned cardiac muscle.11
Another defect in TnI depleted cTnI–/– myocytes was a
reduced responsiveness of the regulatory system to Ca during
the development of active force. This observation is different
from what has been reported in skeletal muscle following
extraction of TnI,23 in which an increase in Ca sensitivity was
observed. This may be an indication that cardiac and skeletal
muscle are fundamentally different in this respect. TnI extraction experiments have been performed in cardiac muscle,
but the effects of TnI removal per se on Ca sensitivity of
tension have not been presented in detail.11,12 We consider it
unlikely that developmental changes in other myofilament
proteins occurring between days 14 and 17 are responsible for
the observed rightward shift in the tension-pCa relationship,
because there was little change in the tension-pCa curve in
wild-type myocytes measured in this time window. Moreover, a candidate myofilament protein has not been identified
in the literature that both undergoes a significant switch
during this brief period and would be able to account for such
a large desensitization of the myofilaments to Ca.
The observed changes in myocyte contractility with TnI
loss are important in view of the possibility that TnI is altered
in certain cardiac-diseased states.4 – 8 In dog myocardium
subjected to coronary occlusion, all 3 troponin subunits were
degraded with TnI showing the most marked change.5 Myofilament Ca sensitivity assessed by superprecipitation of
actomyosin was depressed by this treatment.5 In rat myocardium subjected to 60 minutes of complete global ischemia,
both cTnI and TnT were reduced by '40% to 50%, whereas
other myofilament proteins were unaffected.6 The myofilament Ca sensitivity of actomyosin ATPase activity was
enhanced in this experimental system,6 but this may be due to
binding of cytosolic proteins to the myofilaments.24 In pig
hearts subjected to acute myocardial infarction, cTnI and TnT
levels were reduced by 40% to 80% in myocardial tissue
remote from the infarction zone after 2 months of remodeling.7 In rat experimental myocardial infarction, cTnI content
was reduced by 53% in myocytes remaining 7 days after
surgery.8 In the latter case, assessment of the mechanical
properties of ventricular myocytes revealed an increase in
Huang et al
resting tension and a decrease in Ca sensitivity of active
tension.8 However, substantial increases in phosphorylation
of cTnI and TnT were also observed,8 which could contribute
to the observed decrease in myofilament Ca responsiveness.3,16 In the present study, any influence of TnI phosphorylation by protein kinase A can be ruled out, because the TnI
isoform present in these hearts is lacking protein kinase A
sites. Also, in the present study there was no evidence of new
proteins bound to the myofilaments or of secondary changes
due to fibrosis or tissue remodeling, which can confound
interpretation of mechanical measurements in diseased tissue.
These considerations strengthen our conclusion that TnI
depletion by '50% to 60% is responsible for the mechanical
defects observed in cTnI–/– myocytes including as much as a
2-fold increase in resting tension and a decrease in Ca
sensitivity of tension by more than 0.3 pCa units.
Myocardial stunning is another disease state in which TnI
may be selectively modified. Evidence has been presented in a
rat model of stunned myocardium that Ca-dependent proteolysis
of cTnI, but not other myofilament proteins, underlies the
pathology.4 Physiological measurements in intact ventricular
muscle demonstrated a decreased myofilament Ca responsiveness and altered diastolic tone after stunning.25 A reduction in
myofilament Ca sensitivity has also been observed in stunned
porcine ventricular myocytes.26,27 It should be recognized, however, that changes in TnI in stunned myocardium may not be
analogous to those in the cTnI knockout mice. Proteolysis in
stunned rat myocytes produced a polypeptide fragment of TnI4
that presumably remained associated with the myofilaments;
therefore, this condition cannot be considered equivalent to TnI
depletion. No proteolytic fragments of TnI were detected in the
present study or in other studies reporting reduced TnI levels,7,8
but such fragments could have gone undetected by the antibodies used. A different proteolytic fragment of TnI was reported in
the rat complete global ischemia model,6 but evidence has been
presented that this polypeptide is unrelated to TnI.23 Establishing
the mechanism of ssTnI loss in the knockout mice and the
relationship of this process to cTnI proteolysis in stunned
myocardium and to cTnI loss during ischemic cardiac disease
must await further investigation.
At this point, the cTnI knockout mouse should not be
considered a model of any specific human cardiac disease. It
should nevertheless contribute to our understanding of the
consequences of myocardial TnI deficiency and thereby
facilitate understanding of certain diseased states. Ischemic
heart disease can be associated with diastolic dysfunction and
depressed contractile function,7,8,21,22,28 and the results with
TnI knockout myocytes are consistent with the notion that
TnI modification contributes to the pathology of postischemic
heart disease. Specifically, we propose, on the basis of the
increase in resting tension observed here, that TnI depletion
could contribute to impaired relaxation, increased myocardial
stiffness, and altered ventricular filling during diastole. Moreover, the observed reduction in Ca sensitivity of active
tension would be expected to depress contractility in surviving regions of the beating heart. Indeed, it seems quite
possible that degradation and/or depletion of TnI occurs not
only in necrotic myocardial tissue, but also in myocytes that
survive and contribute to dynamic contractile function.4,7,8
January 8/22, 1999
7
Current evidence argues against the possibility that TnI
depletion is a common defect in heart failure,14 –16,29 although
serum TnI levels can be elevated in severe forms.9 It remains a
formal possibility that TnI insufficiency either due to mutation or
to altered gene expression could be a factor in some forms of
idiopathic cardiomyopathy in humans. The frequency of “severe” TnI mutations in the human population remains to be
established, but several point mutations in TnI have recently
been shown to be associated with some forms of human
hypertrophic cardiomyopathy.30 The line of mice described here
may complement transgenic approaches for creating mice with
interesting TnI mutations (eg, by eliminating complications
arising from endogenous cTnI expression). Moreover, the striking phenotype of the cTnI-null mice, including the sudden onset
and highly predictable time of death, can be exploited as a
bioassay for optimizing delivery of TnI genes, TnI peptides, or
peptidomimetics into myocardium with potentially broad implications for treating cardiac dysfunction.
Targeted ablation of the cTnI gene has shed light on the
mechanism of the TnI isoform switch by showing that the
appearance of cTnI in the developmental program influences
ssTnI expression but is not required for the disappearance of
ssTnI from the heart. These observations suggest that cardiac
expression of the ssTnI gene is “set” in the developmental
program to switch off, and it does so even in the absence of the
adult cTnI isoform. The time course of the TnI isoform switch
was delayed in cTnI1/– heterozygotes and even further delayed
in homozygous cTnI–/– mice. These observations can be rationalized by a model in which a competition exists between TnI
isoforms for a fixed number of binding sites on the thin filament
strand (X.P.H. and J.W.W., unpublished observations, 1998).
TnI molecules that do not find a site are presumably rapidly
degraded by proteolysis. This simple model provides a straightforward explanation for the apparent gene dosage effect on the
time course of the TnI isoform switch in heterozygous cTnI1/–
mice; a reduced amount of cTnI protein would be less effective
competing with ssTnI, resulting in a delayed transition point
during the isoform switch. It also offers an explanation for the
apparent compensation by ssTnI. In this case, the absence of
cTnI prevents it from displacing ssTnI from filament sites.
Eventually, however, ssTnI expression ceases, and when normal
turnover reduces the filament TnI content, a condition of TnI
depletion prevails.
It is remarkable in the case of these cTnI mutant mice
that apparent ssTnI gene inactivation occurs regardless of
the consequences to the animals in terms of survival. This
lack of plasticity in fetal TnI expression in the juvenile
mouse heart is reminiscent of the observation in failing
hearts that the fetal TnI gene, unlike many other contractile
proteins,31,32 fails to become reactivated during the end
stages of heart failure.14 –16,29 Taken together, these observations indicate that ssTnI expression in the mammalian
myocardium is developmentally programmed to switch off
and remain off. Mechanisms responsible for this inactivation of ssTnI expression are currently unclear.
It is worth noting that differences in the behavior of the
thin-filament regulatory system due to the presence of ssTnI
versus cTnI are apparent in the data. At 14 days after birth,
when wild-type myocytes have nearly a full complement of
8
Cardiac Troponin I Knockout
cTnI and null myocytes have nearly a full complement of
ssTnI, a difference of 0.5 pCa units was observed in pCa50
values (Table 2). Thus, myocytes expressing primarily ssTnI
displayed an enhanced Ca sensitivity compared with myocytes expressing primarily cTnI, consistent with previous
reports.1,33 This supports the suggestion that TnI isoforms are
responsible for a large part of the difference in the Ca
responsiveness of neonatal versus adult cardiac muscle.1,33
In summary, the consequences of targeted ablation of the
cTnI gene include partial compensation by altered expression
of ssTnI, followed by a state of TnI depletion when ssTnI
expression declines. Changes in physiological properties that
develop in parallel with TnI depletion include elevated
Ca-independent force, suggesting impaired cardiomyocyte
relaxation, a substantial reduction in the Ca sensitivity of
force development, and ultimately a lethal phenotype due to
acute heart failure. Because loss of ssTnI is rapid and there is
little evidence of fibrosis, infiltration, or hypertrophy, we
conclude that these changes in cardiac function are the result
of TnI depletion. The cTnI knockout mice represent an
important model system for further investigation into the
mechanisms and consequences of TnI depletion and for
controlling the molecular species of TnI to elucidate its role
in cardiac function and in the etiology of cardiac disease.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
Acknowledgments
This work was supported by National Institutes of Health grant P01
HL-47053 (to R.G.G., P.A.P., and J.W.W.). Y.Q.P. was supported by a
postdoctoral fellowship from the American Heart Association, Wisconsin Affiliate, Inc. K.J.L. was supported by a Scientist Development
Grant from the American Heart Association, Inc. J.W.W. was supported
by a National Institutes of Health Research Career Development Award
(K04 HL-03119). The authors wish to thank Drs Gary Lyons, Matthew
Wolff, Timothy Kamp, Marion Greaser, and Richard Moss for helpful
discussions and comments on an earlier version of this manuscript. The
technical assistance of Kara Kemnitz (PCR), Inge Sigglekow (heart
tissue sections), and Dr Robert Conhaim (lung tissue sections and
analysis) is gratefully acknowledged.
References
1. Solaro RJ, Van Eyk J. Altered interactions among thin filament proteins
modulate cardiac function. J Mol Cell Cardiol. 1996;28:217–230.
2. Solaro RJ, Moir AJG, Perry SV. Phosphorylation of troponin I and the
inotropic effect of adrenaline in perfused rabbit heart. Nature. 1976;262:
615– 617.
3. Holroyde MJ, Howe E, Solaro RJ. Modification of calcium requirements
for activation of cardiac myofibrillar ATPase by cyclic AMP dependent
phosphorylation. Biochem Biophys Acta. 1979;586:63– 69.
4. Gao WD, Atar D, Liu Y, Perez NG, Murphy AM, Marban E. Role of
troponin I proteolysis in the pathogenesis of stunned myocardium. Circ
Res. 1997;80:393–399.
5. Toyo-Oka T, Ross J. Ca sensitivity change and troponin loss in cardiac
natural actomyosin after coronary occlusion. Am J Physiol. 1981;240:
H704 –H708.
6. Westfall MV, Solaro RJ. Alterations in myofibrillar function and protein
profiles after complete global ischemia in rat hearts. Circ Res. 1992;70:
302–313.
7. Ricchuiti V, Zhang J, Apple FS. Cardiac troponin I and T alterations in hearts
with severe left ventricular remodeling. Clin Chem. 1997;43:990–995.
8. Li P, Hofmann PA, Malhorta A, Cheng W, Sonnenblick EH, Meggs LG,
Anversa P. Myocardial infarction alters myofilament Ca sensitivity and
mechanical behavior of myocytes. Am J Physiol. 1997;272:H360 –H370.
9. Missov E, Calzolari C, Pau B. Circulating cardiac troponin I in severe
congestive heart failure. Circulation. 1997;96:2953–2958.
10. Antman EM, Tanasijevic MJ, Thompson B, Schactman M, McCabe CH,
Cannon CP, Fischer GA, Fung AY, Thompson C, Wybenga D, Braunwald E.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
Cardiac specific troponin I levels to predict the risk of mortality in patients
with acute coronary syndromes. N Engl J Med. 1996;335:1342–1349.
Strauss JD, Zeugner C, Van Eyk JE, Bletz C, Troschka M, Ruegg JC.
Troponin replacement in permeabilized cardiac muscle: reversible extraction
of troponin I by incubation with vanadate. FEBS Lett. 1992;310:229–232.
Strauss JD, Van Eyk JE, Barth Z, Kluwe L, Wiesner RJ, Maeda K, Ruegg
JC. Recombinant troponin I substitution and calcium responsiveness in
skinned cardiac muscle. Pflügers Arch. 1996;431:853– 862.
Schianffino S, Gorza L, Ausoni S. Troponin isoform switching in the
developing heart and its functional consequences. Trends Cardiovasc
Med. 1993;3:12–17.
Hunkeler NM, Kullman J, Murphy AM. Troponin I isoform expression in
human heart. Circ Res. 1991;69:1409 –1414.
Sasse S, Brand NJ, Kyprianou P, Dhoot GK, Wade R, Arai M, Periasamy M,
Yacoub MH, Barton JR. Troponin I gene expression during human cardiac
development and in end-stage heart failure. Circ Res. 1993;72:932–938.
Bodor GS, Oakley AE, Allen PD, Crimmins DL, Ladenson JH, Anderson,
PA. Troponin I phosphorylation in the normal and failing adult human
heart. Circulation. 1997;96:1495–1500.
Gomez AM, Valdivia HH, Cheng H, Lederer MR, Santana LF, Cannell
MB, McCune SA, Altschuld RA, Lederer WJ. Defective excitationcontraction coupling in experimental cardiac hypertrophy and heart
failure. Science. 1997;276:800 – 806.
Gregg RG, Messing A, Strube C, Beurg M, Moss R, Behan M, Sukhareva
M, Haynes S, Powell JA, Coronado R, Powers PA. Absence of the
b-subunits of skeletal muscle dihydropyridine receptor alters expression
of the a1 subunit and eliminates excitation-contraction coupling. Proc
Natl Acad Sci U S A. 1996;93:13961–13966.
Huang XP, Pi YQ, Lokuta AJ, Greaser M, Walker JW. Arachidonic acid
stimulates protein kinase C epsilon redistribution in heart cells. J Cell Sci.
1997;110:1625–1634.
Araujo A, Walker JW. Kinetics of tension development in skinned
cardiac myocytes measured by photorelease of calcium. Am J Physiol.
1994;267:H1643–H1653.
Pfeffer MA, Braunwald E. Ventricular remodeling after myocardial
infarction. Circulation. 1990;81:1161–1172.
Colucci WS. Molecular and cellular mechanisms of myocardial failure.
Am J Cardiol. 1997;80:15L–25L.
Shiraishi F, Yamamoto K. The effect of partial removal of troponin I and
C on the Ca sensitive ATPase activity of rabbit skeletal muscle.
J Biochem. 1994;115:171–173.
Barbato, R, Menabo R, Carafoli E, Schiaffino S, Di Lisa F. Binding of
cytosolic proteins to myofibrils in ischemic rat hearts. Circ Res. 1996;
78:821– 828.
Gao WD, Atar D, Backx PH, Marban E. Relationship between intracellular calcium and contractile force in stunned myocardium. Circ Res.
1995;76:1036 –1048.
Hofmann PA, Miller WP, Moss RL. Altered calcium sensitivity of isometric tension in myocyte-sized preparations of porcine postischemic
stunned myocardium. Circ Res. 1993;72:50 –56.
MacDonald K, Mammen PPA, Strang KT, Moss RL, Miller WP. Isometric and dynamic contractile properties of porcine skinned cardiac
myocytes after stunning. Circ Res. 1995;77:964 –972.
Kusuoka H, Marban E. Cellular mechanisms of myocardial stunning.
Annu Rev Physiol. 1992;54:243–246.
Cummins DV, Seymour AM, Rix LK, Kellett R, Dhoot GK, Yacoud MN,
Barton PJ. Troponin I, T protein expression in experimental cardiac
hypertrophy. Cardioscience. 1995;6:65–70.
Kimuara A, Harada H, Park E-J, Nishi H, Satoh M, Nakahashi M, Hiroi
S, Sasaoka T, Ohbuchi N, Nakamura T, Koyanagi T, Hwang T-H, Choo
J-A, Chung K-S, Hasegawa A, Nagai R, Okazaki O, Nakamura H,
Matsuzaki M, Sakamoto T, Toshima H, Koga Y, Imaizumi T, Saszuki T.
Mutations in the cardiac troponin I gene associated with hypertrophic
cardiomyopathy. Nat Genet. 1997;16:379 –382.
Boheler KR, Schwartz K. Gene expression in cardiac hypertrophy. Trends
Cardiovasc. 1992;2:176 –182.
Molkenten JD, Lu J-R, Antos CL, Markham B, Richardson J, Robbins J,
Grant SR, Olson ER. A calcineurin-dependent transcriptional pathway for
cardiac hypertrophy. Cell. 1998;93:215–228.
Westfall MV, Rust EM, Metzger JM. Slow skeletal troponin I gene
transfer, expression and myofilament incorporation enhances adult
cardiac myocyte contractile function. Proc Natl Acad Sci U S A. 1997;
94:5444 –5449.
Fish D, Orenstein J, Bloom S. Passive stiffness of isolated cardiac and
skeletal myocytes in the hamster. Circ Res. 1984;54:267–276.