Despite a history dating back to the 1800s, using Clostridium bacteria to treat cancer has not advanced beyond the observation that they can colonise and partially destroy solid tumours. Progress has been hampered by their inability to eradicate the viable portion of tumours, and an instinctive anxiety around injecting patients with a bacterium whose close relatives cause tetanus and botulism. However, recent advances in techniques to genetically engineer Clostridium species gives cause to revisit this concept. This paper illustrates these developments through the attenuation of C. sporogenes to enhance its clinical safety, and through the expression and secretion of an immunotherapeutic. An 8.6 kb sequence, corresponding to a haemolysin operon, was deleted from the genome and replaced with a short non-coding sequence. The resultant phenotype of this strain, named C. sporogenes-NT, showed a reduction of haemolysis to levels similar to the probiotic strain, C. butyricum M588. Comparison to the parental strain showed no change in growth or sporulation. Following injection of tumour-bearing mice with purified spores of the attenuated strain, high levels of germination were detected in all tumours. Very low levels of spores and vegetative cells were detected in the spleen and lymph nodes. The new strain was transformed with four different murine IL-2-expressing plasmids, differentiated by promoter and signal peptide sequences. Biologically active mIL-2, recovered from the extracellular fraction of bacterial cultures, was shown to stimulate proliferation of T cells. With this investigation we propose a new, safer candidate for intratumoral delivery of cancer immunotherapeutics.