J Mol Cell Cardiol 29, 1499–1504 (1997)

Direct Gene Transfer into the MouseHeartKai Li, Robert E. Welikson, Karen L. Vikstrom and Leslie A. LeinwandDepartment of Microbiology and Immunology, Albert Einstein College of Medicine, NY, and theDepartment of Molecular Cellular Developmental Biology, University of Colorado, CO, USA

(Received 30 May 1996, accepted in revised form 21 January 1997)

K. L, R. E. W, K. L. V L. A. L. Direct Gene Transfer into the Mouse Heart. Journalof Molecular and Cellular Cardiology (1997) 29, 1499–1504. Direct injection of plasmid DNA into the myocardiumof several species has been shown to be useful for studying cardiac gene expression. However, despite a betterunderstanding of mouse genetics and the availability of several disease models in mice, gene injection withplasmid DNA into the mouse heart has not been reported. In this study, we demonstrate a simple and reproduciblemethod for gene transfer into the mouse heart via direct injection of plasmid DNA. A firefly luciferase gene,driven by the RSV promoter, was used to quantitatively determine the spatial and temporal characteristics ofgene transfer. Luciferase gene expression was stable for 8 weeks and showed a dose-dependent response over arange of 0.3–3 lg of input DNA. Inter-animal variability was low and gene expression was restricted to the leftventricle, near the site of injection. This method was also demonstrated to be suitable for detecting the expressionof structural genes under the control of cellular promoters. Immunohistochemistry was used to detect theexpression of an epitope-tagged myosin heavy chain driven by a rat a-myosin heavy chain promoter. Thus,naked DNA injection into the mouse heart results in a highly reproducible expression of constructs with eitherviral or cellular promoters. It is a relatively inexpensive and efficient means of studying cardiac gene regulationin vivo and a useful tool for screening the potential transgenes before generating transgenic mice.

1997 Academic Press Limited

K W: Gene expression; Mouse heart; Gene injection.

efficient, and results in up to 50-fold higher levelsIntroductionof gene expression than similar experiments inskeletal muscles (Lin et al., 1990; Kitsis et al., 1991;Since the initial demonstration of direct gene in-

jection into mouse skeletal muscle (Wolff et al., Buttrick et al., 1992).Transfer of naked plasmid DNA into the heart1990), gene injection into various skeletal muscle

sites has been studied, and various parameters has subsequently been demonstrated in rabbits anddogs (Gal et al., 1993; von Harsdorf et al., 1993).for optimising gene transfer have been tested. A

mouse’s hind limbs, diaphragm, and tongue can Although the mechanism underlying the uptake ofinjected genes remains unknown, it appears thateach take up and express injected genes (Acsadi et

al., 1991a; Wolff et al., 1991; 1992; Davis et al., this ability is restricted to striated muscles. In ratand rabbit hearts, gene expression after a single1993). Gene transfer into a normal adult skeletal

muscle of mouse and rat is quite inefficient, al- DNA injection is generally stable over time. Thepotential applications of direct gene transfer includethough agents which promote muscle de-

generation/regeneration dramatically increase the potential somatic gene therapy, the analysis of generegulation in vivo (Kitsis et al., 1991; Buttrick et al.,expression of injected genes (Davis and Jasmin,

1993; Danko et al., 1994). In contrast to skeletal 1992), vaccination (Raz et al., 1994), and hormonereplacement therapy (Leinwand and Leiden, 1991).muscle, gene injection into the rat heart is quite

Please address all correspondence to: Leslie A. Leinwand, Department of Molecular Cellular Developmental Biology, Campus Box 347,University of Colorado, Boulder, CO 80309-0347, USA.

0022–2828/97/051499+06 $25.00/0 mc970389 1997 Academic Press Limited

K. Li et al.1500

Despite the prevalent use of the mouse in genetic chain (aMyh) driven by a 2.9 kb rat aMyh promoter(Vikstrom and Leinwand, unpubl) was also used.and developmental studies, there are no published

reports of DNA injections into the mouse heart. We For luciferase assays, mice were killed by cervicaldislocation 1, 2, 4, and 8 weeks after gene injection.have developed surgical protocols for mouse cardiac

injection and characterized the resulting gene ex- The hearts were removed and washed with ice-cold homogenization buffer (25 m Gly-gly, 15 mpression. Our data show that gene injection into

the mouse heart is highly reproducible, and results MgSO4, 4 m EGTA, and 4 m DTT). The tissuewas minced in 0.4 ml homogenization buffer,in stable gene expression over 8 weeks. When

reporter genes were used, expression was dose- homogenized for 20 s with a Tissuemizer (TekmarCompany, Cincinnati, OH, USA), and then cent-dependent and restricted to a limited area of the left

ventricle. These results are consistent with previous rifuged at 5000×g for 20 min (4°C). The super-natant was removed and kept for luciferase assays.reports of gene injection in rat (Buttrick et al., 1992)

and dog hearts (von Harsdorf et al., 1993). The Twenty-five percent of the supernatant (100 ll) wasmixed with 360 ll reaction buffer (25 m Gly-gly,detection of a tagged mammalian contractile pro-

tein gene under the control of a cellular promoter, 15 m MgSO4, 4 m EGTA, 4 m DTT, 15 m

KPO4, pH 7.8, and 2 m Na+-ATP). The enzymaticsuggests that it may also be possible to evaluatecandidate transgenes with this method. The con- reaction was initiated by injection of 100 ll

homogenization buffer containing 0.2 m D-lu-sistency of the results obtained with gene injectioninto the mouse heart has opened a new avenue for ciferin and light emission was measured for 20 sec

(Brasier et al., 1989). The light units measuredstudying mouse cardiac gene expression in vivo,using the most prevalent species in mammalian from the 100 ll of supernatant were expressed as

relative luciferase activity.genetic analysis.For tissue sectioning and immunostaining, the

heart was harvested 5 days after gene injection,flash frozen in liquid nitrogen and stored atMaterials and Methods−80°C. After embedding in OCT, 5 lm sectionswere collected every 20 sections and placed onAdult female CD1 mice were anesthetized with

chloral hydrate (0.4 g/kg, i.p.). The animals were poly--lysine coated slides. The slides were air-dried, fixed in acetone (−20°C for 5 min), and thenplaced in a supine position and the upper limbs

were taped to a table. Chest skin was cleaned with allowed to air dry again. Endogenous peroxidaseactivity was blocked by incubating in 0.3%70% ethanol and a 1–1.5 cm incision was made

along the left side of the sternum. The muscle layers hydrogen peroxide in methanol for 30 min atroom temperature. All subsequent incubationsof the chest wall were bluntly dissected to avoid

bleeding. The thorax was opened by cutting the were done at 37°C in a humidified chamber.After air drying, the slides were blocked byrib at the point of the most pronounced cardiac

pulsation. While forceps were used to widen the incubating for 30 min with 10% goat serum inphosphate buffered saline (PBS). The sections werechest, the abdomen and the right side of the chest

were pressed to push the heart out of the thoracic then incubated with 9E10.2 culture supernatant(Bishop and Evan, 1985; ATCC CRL 1729) forcavity. Unless specified, 5 lg of DNA in 15 ll normal

saline was injected through a 30-gauge needle into 30 min. The slides were washed by soaking inthree, 5 min changes of PBS and then incubatedthe left ventricular free wall. Following the injection,

the heart was placed back into the thoracic cavity for 30 min with a 1:200 dilution (in 10% goatserum) of horseradish peroxidase conjugated goatand the chest held closed with forceps. After the

muscle layers were closed with one suture, the anti-mouse IgG (Bio-Rad, Hercules, CA, USA).Slides were washed as above, and color wasforceps were released. The chest cavity was gently

squeezed to expel air out before tightly tying the developed using the DAB substrate kit (VectorLaboratories, Inc., California, CA, USA), accordingsuture and stapling the skin. Spontaneous res-

piration was maintained and no mechanical vent- to the manufacturer’s instructions. Slides weremounted with Airvol (Air Products & Chemicalilation was needed.

Two plasmids were used for this study. The re- Inc., Allentown, PA, USA).All results are expressed as mean±... A one-porter gene, pRSVluc (Buttrick et al., 1992) contains

a firefly luciferase cDNA fused to the long terminal way analysis of variance followed by a Student–Newman–Keuls test was used for multiple com-repeat of the Rous Sarcoma Virus (RSV LTR) (Gor-

man et al., 1982). A candidate transgene paa-myc, parisons. A P value less than 0.05 was used asstatistical significance.encoding a myc epitope-tagged rat a-myosin heavy

Direct Gene Transfer into the Mouse Heart 1501

8

010

Dosage of DNA injected ( g)

Rel

ativ

e L

UC

act

ivit

y (l

og)

6

4

2

310.3

*

7

Time after injection (weeks)

Rel

ativ

e L

UC

act

ivit

y (l

og)

5

6

4

3

2

1

2 4 6 80

*

Figure 1 Relative luciferase (LUC) activity measured in Figure 2 The level of luciferase (LUC) gene expressionmouse myocardial extracts as a function of DNA dose as a function of time, where n=6 for each group, exceptinjected. All values of relative luciferase activity were n=7 for the 8 week group. Gene expression was stableobtained from 100 ll supernatant. Values are the 8 weeks examined. ∗=P<0.05 as compared to themean±... ∗=P<0.05 as compared to the next group other groups.with smaller dose.

Results

The survival of mice from the surgery and geneinjection was quite high (76%, n=55). After geneinjection, death resulting from depressed respirationoccurred primarily within the first h. To maximisesimplicity, no respirator was used for this surgery.With the luciferase reporter gene, high levels ofluciferase activity were detected in all surviving

7

0LV

Space distribution of reporter gene

Rel

ativ

e L

UC

act

ivit

y (l

og)

5

6

4

3

2

1

SeptumA + RV

mice using doses of DNA from 0.3 to 10 lg. With theexception of one heart, which probably represents a Figure 3 Spatial distribution of luciferase (LUC) ex-poor gene injection, the inter-animal variability pression after a single injection into the mouse heart.

Three hearts were included in this group and all showedin luciferase activity was less than one order ofhigh levels of expression in the left ventricle (LV) and nomagnitude within any group. To determine thedetectable luciferase activity in atrium (A), right ventricleoptimal DNA dose for gene injection into the mouse (RV), and septum.

heart for the RSV-luciferase construct, four dosesof pRSVLuc were tested: 0.3, 1, 3, and 10 lg, allin a volume of 15 ll. Each dose of DNA was injected injected mouse hearts was assessed using the luci-into five mice, and luciferase activity was assayed ferase reporter gene. Luciferase assays were per-1 week later. As shown in Figure 1, luciferase formed on dissected regions of the heart (n=3). Asactivity increased in a dose-responsive manner with shown in Figure 3, only the left ventricle expressedinjections of up to 3 lg DNA, then plateaued. This the reporter gene. Since reporter gene assays, suchsuggests that to achieve gene expression within the as luciferase activity, are extremely sensitive, welinear range of a RSV LTR driven construct, 3 lg wished to assess whether expression of genes drivenof DNA is the optimal dose. by cellular promoters can be detected following

By measuring luciferase levels at various times gene injection into the mouse heart. Figure 4 showspost-injection, we addressed whether the expression the distribution of myc-tagged aMyh protein 1 weekof a reporter gene was stable over time (Fig. 2). after injection of an aMyh-promoter-driven con-The highest level of reporter gene expression was struct into the mouse heart. Immunostaining withdetected 1 week after gene injection (P<0.05 as a myc-specific monoclonal antibody only revealedcompared to the other three groups). The expression gene expression around the needle track in the leftlevel decreased slightly thereafter, but remained ventricle (Fig. 4A). Many cardiac myocytes werestable until 8 weeks post injection. positive for the myc-tagged Myh protein in this

region. However, there were negative cells withinThe spatial distribution of gene expression in

K. Li et al.1502

heart, thus, provides an easy and useful way fordirect gene transfer into a widely used animalspecies.

Gene injection into the mouse heart is fast andsimple. Neither extensive sterile technique nor ar-tificial respiration is required. A 77% survival rateand a 100% incidence of positive expression dem-onstrate its reliability and reproducibility. Any vari-ability of gene expression in the hearts of smallmammals, such as the rat and the mouse, is prob-ably due to the quality of the injection. The smallsize and thin free wall of the mouse heart maysignificantly contribute to the variability in thisanimal. Our previous gene injection studies in therat heart included a vital dye in the DNA solutionto guide the injection (Buttrick et al., 1992). In thisstudy, no dye was used since a limited injectionvolume was allowed. However, we observed evidentsubepicardial edema immediately following theDNA injection which caused the tissue to changecolor from red to white. If bleeding was avoidedand the chest cavity was not open for longer than2 min, the survival rate was quite high.

From several standpoints, the behavior of genesinjected into the mouse heart is similar to thatFigure 4 Immunostaining for myc tagged rat a myosin

heavy chain expression in the mouse heart 1 week after observed in the hearts of rats and dogs (Lin et al.,a 5 lg DNA injection of paamyc plasmid. Dark brown 1990; Acsadi et al., 1991a; Buttrick et al., 1992;staining indicates myc tagged rat a myosin. An injected von Harsdorf et al., 1993). Our luciferase dataheart is shown in (A) and an uninjected is shown in (B).

indicate that only the left ventricle expressed theBar equals 20 lm.injected reporter gene, and the results from paa-myc injection further showed that expression of theinjected gene was largely restricted to the injectionthe needle track as well. The highest density ofsite vicinity. Expression from the RSV-LTR drivenpositive cells was seen in the central part of theconstruct behaved in a dose-dependent manner inneedle track with fewer positive cells around thethe range of 0.3–3 lg DNA. However, given theperiphery of the track. There were no positive cellsquantitatively different promoter activities existingwith the myc antibody in an uninjected mousein mammalian cells, it is important to define theheart (Fig. 4B).optimal dosage for each expression construct in-jected.

Gene expression after plasmid DNA injection isgenerally stable over time, which is one of theDiscussionadvantages of this method when compared to themore transient gene expression following re-Three observations result from this study. Firstly,

we have demonstrated the feasibility of direct gene combinant adenovirus injection (for review seeLeinwand and Leiden, 1991; Nadal-Ginard andtransfer into the mouse heart. Secondly, we have

defined the behavior of genes injected into the Mahdavi, 1993). Finding peak gene expression level1 week post-injection in the mouse heart is con-mouse heart with respect to the spatial and temporal

distribution of gene expression, and thus, provided sistent with the findings in rats and dogs (Lin etal., 1990; Kitsis et al., 1991; Buttrick et al., 1992;a range of doses at which a high level expression

of injected genes can be obtained. The third con- von Harsdorf et al., 1993). The duration of geneexpression in the mouse heart is stable over 8tribution of this study, is in demonstrating the

ability to evaluate potential transgenes in vivo before weeks. The result is similar to findings from previousstudies in the rat, showing gene expression for upproducing a transgenic line. This finding is notable,

given the inefficiency of transfection into cultured to 2 months following a single injection (Lin et al.,1990; Kitsis et al., 1991; Buttrick et al., 1992). Incardiac myocytes. Gene injection into the mouse

Direct Gene Transfer into the Mouse Heart 1503

addition, very stable gene expression, even over Acknowledgement1 year, in mouse skeletal muscles has been dem-onstrated after a single plasmid DNA injection (Wolff The authors thank Drs Kass-Eisler, Buttrick and

Geenen for their help with the pilot study, andet al., 1992). Whether gene expression in the mouseheart after DNA injection can persist for the same Drs Kass-Eisler, Roopnarine and Buttrick for their

critical reading of this manuscript. KL is a recipienttime period is not addressed in this study, but therelatively stable gene expression we observed over of a postdoctoral fellowship of the American Heart

Association New York City Affiliate, and LAL is a8 weeks suggests that it may be possible.Compared to other species, direct DNA injection recipient of NIH grant 5R37HL50560-03.

into the mouse heart has several obvious ad-vantages for molecular genetic studies. First, thegenetics of the mouse are well understood, and Referencesgenetic manipulations in the mouse are nowperformed relatively routinely. A major con- A G, J S, J A, D D, W P, W

W, W JA, 1991a. Direct gene transfer and ex-tribution of DNA injection into animal hearts ispression into rat heart in vivo. New Biol 3: 71–81.in the in vivo analysis of gene regulation. An

A G, D G, L DR, J A, W FS,increasing number of cis- and trans-activatingG A, W JA, D KE, 1991b. Human

factors have been identified for mammalian genes, dystrophin expression in mdx mice after intramusculargenerally using transfection of cultured cells (Mor- injection of DNA constructs. Nature 352: 815–818.

B JM, E GI, 1985. Isolation of monoclonal anti-kin, 1993). Such approaches for studying cardiacbodies specific for human c-myc protooncogene prod-gene regulation have been limited by a lack ofuct. Mol Cell Biol 5: 3610–3616.permanent cardiac cell lines and an inability to

B AR, T JE, H JF, 1989. Optimized usetransfect adult primary cardiac myocytes in cul- of the luciferase assay as a reporter gene in mammalianture. DNA injection into the mouse heart should cell lines. Biotechniques 7: 1116–1122.

B PM, K A, K RN, K ML, Lbecome very useful for the study of cardiac geneLA, 1992. Behavior of genes directly injected into theregulation in vivo, as it has been in the rat (Kitsisrat heart in vivo. Circ Res 70: 193–198.et al., 1991; Buttrick et al., 1992). Second,

D I, F JD, J S, H K, L JS,many mouse models for human diseases can be W JA, 1994. Pharmacological enhancement of insubstrates for testing gene therapeutic in- vivo foreign gene expression in muscle. Gene Ther 1:

114–121.terventions. One example is the mdx mouse,D HL, J BJ, 1993. Direct gene transfer intowhich is deficient in dystrophin, and has been

mouse diaphragm. FEBS Let 333: 146–150.used to test several approaches in the treatmentD HL, D B, Q B, C J, W

of muscular dystrophies. These approaches include R, 1993. Plasmid DNA is superior to viral vectors forgene transfer by either direct DNA injection direct gene transfer into adult mouse skeletal muscle.

Hum Gene Ther 4: 733–740.(Acsadi et al., 1991b) or viral vector infectionG D, W L, L G, P JG, H J, I(Karpati and Acsadi, 1994). In addition, following

JM, 1993. Direct myocardial transfection in two animaldirect DNA injection of a tRNA suppressor genemodels. Evaluation of parameters affecting gene ex-

into the mdx mouse heart, we recently observed pression and percutaneous gene delivery. Lab Investa phenotypic correction of cardiac myocytes (Li 68: 18–25.

G CM, M GT, W MC, P I,and Leinwand, in preparation). Since the myo-H BH, 1982. The Rous sarcoma virus longcardium takes up DNA more efficiently than otherterminal repeat is a strong promoter when introducedtissue types, gene injection into the mouse heartinto a variety of eukaryotic cells by DNA mediated

provides a simple way to assay potential gene transfection. Proc Natl Acad Sci USA 79: 6777–6781.therapies in mouse models of genetic diseases. K G, A G, 1994. The principles of gene

therapy in Duchenne muscular dystrophy. Clin InvestIn conclusion, this study demonstrates the feas-Med 17: 499–509.ibility of direct DNA injection into the mouse heart

K R, B P, MN E, K M, Land further characterizes the behavior of injectedL, 1991. Hormonal modulation of a gene injected

genes. Clearly, the mouse heart efficiently takes up into rat heart in vivo. Proc Natl Acad Sci USA 88:and expresses injected genes in a dose-dependent 4138–4124.

L LA, L JM, 1991. Gene transfer into car-pattern. Expression of injected naked DNA in thediac myocytes in vivo. Trends Cardiovasc Med 1: 271–mouse heart is restricted in the localized area of276.injection and is relatively stable over time. Gene

L H, P MS, M G, B S, L JM,injection into the mouse heart provides a highly 1990. Expression of recombinant genes in myocardiumefficient and reproducible tool for molecular genetic in vivo after direct injection of DNA. Circulation 82:

2217–2221.studies in this well-defined species.

K. Li et al.1504

M E, 1993. Regulation of myosin heavy chain genes M V, N-G B, 1993. Gene injectioninto canine myocardium as a useful model for studyingin the heart. Circulation 87: 1451–1460.

N-G B, M V, 1993. Molecular mech- gene expression in the heart of large mammals. CircRes 72: 688–695.anisms of cardiac gene expression. Basic Res Cardiol

88 (Suppl 1): 65–79. W JA, M RW, W P, C W, AG, J A, F PL, 1990. Direct gene transfer intoR E, C DA, P SE, P TB, A AM,

A G, G SH, S M, L D, mouse muscle in vivo. Science 247: 1465–1468.W JA, W P, A G, J S, J A, CY MA, B SM, R GH, 1994.

Intradermal gene immunization: the possible role W, 1991. Conditions affecting direct gene transfer intorodent muscle in vivo. Biotechniques 11: 474–485.of DNA uptake in the induction of cellular

immunity to viruses. Proc Natl Acad Sci USA 91: 9519– W JA, C W, A G, W P, 1992.Expression of genes injected into mouse muscle in vivo.9523.

H R, S RJ, S YT, V SF, Human Mol Genet 1: 53–63.