We, and other have previously shown that using optimised formulations adenovirus (Ad) is several logs more efficient in myocardial gene transfer than “naked” plasmid DNA (pDNA). Here, we evaluated “magnetofection”, the use of magnetic force acting on vectors complexed to magnetic particles and ultrasound to increase non-viral gene transfer to the murine myocardium. C57BL6/J mice were intubated and a thoracotomy performed to expose the heart. The ability of the magnetic force to pass through tissue was tested ex vivo on a dissected lamb's heart. Iron filings sprinkled on the endocardial site moved and clustered together when the magnet was placed on the epicardial side of the left myocardial wall (n=3). A variety of magnetic particles have been described, here we used TransMAGPEI (Chemicell, Germany), a polyethyleneimine (PEI)–coated iron oxide particle. For the myocardial injection the total volume could not exceed 10 μl. The magnetic particle:DNA complexes were carefully optimised by altering the TransMAGPEI to pDNA w/w and v/v ratios to avoid precipitation. The complexes used for in vivo injections were made in the following way: 0.8 μl TransMAGPEI (60 μg/μl) was diluted with 41.7 μl water and mixed with 5 μl pDNA (5 μg/μl) carrying a luciferase reporter gene (pLux). 2.5 μl 0.9% NaCl (18%) was added after 15 minutes. 10 μl of this formulation containing 5 μg pLux were injected into the left ventricular wall. The hearts were exposed to the magnet (AlNiCo Rod magnet plus NeoDelta magnet, IBS) for 5 min straight after injection and for a further 10 min after re-sutering of the chest. The left ventricular wall was harvested 48 hrs after gene transfer and lux expression was measured. Although magnet exposed hearts expressed more lux than the non-magnet exposed heart, this difference was not significant (magnet: 268±124 RLU/mg protein, no magnet: 94.0±41.8 RLU/mg protein, n=8-10/group). It has previously been shown that ultrasound (US) enhances non-viral gene transfer in skeletal muscle probably through acoustic cavitation of the cell membrane. Here, US was applied for 5 minutes before and after myocardial injection of 10 μg pLux in 10 μl total volume containg 1:1 v/v Opitson microbubbles. For this purpose the chest cavity was filled with gel (Henleys Medical, Welwyn Garden City, UK) and a 0.5 cm2 probe massaged over the heart area. The Medi-Link system machine (Electro-Medical Supplies Ltd, Wantage, UK) was set to deliver 1MHz, 2 W/cm2 at 100% duty cycle continous wave. 48 hrs after gene transfer the left vetricular wall was harvested and lux expression measured. In a parallel experiment we demonstrated that ultrasound imaging (14 MHz) using a gel filled chest cavity is feasible and therefore concluded that ultrasound waves were reaching the heart. However, the US conditions described above did not increase lux expression in the mouse heart (US: 71.8±38.4 RLU/mg protein, no US:138±69.2 RLU/mg protein, n=10/group). In conclusion, although physical energy may be useful in improving non-viral gene transfer to the myocardium further optimisation of these techniques will be required.