Article Text

Gene therapy made difficult
  1. P A KINGSTON, Bristol-Myers Squibb Cardiovascular Research Fellow
  1. A M HEAGERTY, Professor of Medicine, University of Manchester
  1. Molecular Medicine Unit, Department of Medicine, University of Manchester
  2. Oxford Road, Manchester M13 9PT, UK
  3. email: MDMMSPAK{at}

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Editor,—While we found your recent editorial on gene therapy very interesting,1 some points were raised that invite further comment.

Inflammatory responses seem inevitable following exposure to “first generation” adenovirus vectors; however, transgene selection appears to be an important factor in avoidance of these responses.

Your editorial states “ . . .(t)his inflammatory response is . . .generally observed using the sort of adenoviral loads needed to achieve expression of the transgene”. Undoubtedly, many early in vivo studies of adenovirus mediated gene therapy required very high virus doses to elicit significant transgene expression and therapeutic effects. However, a number of recent studies have obtained significant results with much lower virus doses. Sataet al, using an adenovirus expressing Fas-ligand (a cell surface/secreted protein), achieved a significant reduction in neointima formation with a dose of 1 × 106plaque forming units (pfu)—approximately 1000-fold lower than doses typically used in trials of cytostatic treatment.2 Shearset al demonstrated reduced neointima formation using an iNOS expressing vector at a similarly low virus dose (2 × 106 pfu).3 Therefore, in vascular tissues, transgenes giving rise to either a secreted protein or a protein that gives rise to a secreted product seem to afford some advantage, perhaps by requiring infection of only a small percentage of cells in the vessel wall. In both studies, transgene expression was under the control of the cytomegalovirus immediate–early promoter. It is probable that the use of smooth muscle cell specific promoters (in the vascular setting) will allow more efficient transgene expression and therapeutic effects from even lower virus doses with concomitantly reduced inflammatory responses.

As your editorial suggests, injudicious use of non-autologous transgenes may result in transgene induced immune responses. However, both autologous and non-autologous transgenes, which themselves downregulate the host immune responses to vector administration, have been shown to improve substantially transgene expression and persistence.2 ,4

Contrary to Dr Clesham’s suggestion, deletion of adenoviral genes from vectors has offered a substantial—if not quite revolutionary—improvement in vector efficiency. Stable transgene expression has been demonstrated in immunocompetent mice 10 months after a single injection of “gutless” adenovirus vector expressing α1-antitrypsin from genomic DNA.5Furthermore, “gutless” vectors with space for the insertion of 30 kb of DNA allow the prospect of efficient transgene expression from genomic DNA and production of vectors containing a variety of transgenes, some of which may be aimed at suppression of the host immune response to the vector.

Finally, while host inflammatory responses have attracted much attention, their practical sequelae are not clearly defined in vascular tissues. Despite evidence suggesting that the inflammatory responses in intact arteries may cause neointimal hyperplasia,6 all studies of adenovirus mediated vascular gene therapy that have compared “no virus” and “control virus” groups have demonstrated no significant difference. Inflammatory changes are undeniably precipitated by exposure to adenovirus vectors, but they do not appear to be deleterious in the setting of gene therapy for restenosis.

It is wise to exercise caution regarding the prospects of human gene therapy, but the omens are less portentous than Dr Clesham suggests. Many of the technical problems initially encountered have been addressed successfully, while rapid progress is being made in others. Much of the future difficulty for gene therapy lies in determining which genes offer the best prospects as therapeutic agents rather than in struggling to make poorly expressed, pro-inflammatory transgene products fit roles to which they are not suited. While there is little virtue in pressing ahead recklessly with what are still largely experimental treatments, it seems unlikely that we will have to wait 25 years before the first human is successfully treated by direct gene transfer, particularly in the vascular setting.


This letter was shown to the author, who replies as follows:

Kingston and Heagerty raise a number of important issues in their response to my editorial on gene therapy and I am grateful for their interest.

While the prospects for this emerging technology are unknown, some more definite conclusions can be drawn from the past 10 years. It should be remembered that there is no gene therapy in clinical use at present, despite an almost unprecedented research effort.

One of the underlying aims of the current approach is that the biological effects observed following gene transfer should result from the expressed transgene rather than the vector that delivers it. Unfortunately, acute inflammation at the site of vector delivery is an inevitable, non-specific response to conventional doses of adenoviral vectors. This inflammatory response appears to be independent of the transgene or native adenoviral gene expression as ultraviolet inactivated or defective adenoviral particles can induce inflammation and activate the transcription factor NFkB1-1 1-2 This side effect is particularly unhelpful in the context of arterial gene therapy given the current understanding of atherosclerosis as an inflammatory disorder.1-3

Given the non-specific effects of high adenoviral loads, the ability to use very low adenoviral doses seems attractive and may be possible if more potent promoters are incorporated into gene transfer vectors.1-4 Reports describing the use of very low viral loads of cytomegalovirus driven vectors (106 pfu) are inconsistent with the findings of the vast majority of researchers in this field. Gene transfer in vivo is an inherently inefficient process; there are few reports of meaningful dose–response curves and even fewer examples of excessive transgene expression.

The immune response to adenovirus mediated gene transfer has been extensively studied and has driven the development of so called “gutless” vectors. These “ultimate” adenoviral vectors have been around for some years1-5; however, I am unaware of any significant impact of these newer adenoviral vectors on the disappointing results of hundreds of human gene therapy protocols over the past decade.

Careful evaluation of the problems of inefficiency, local inflammation, and regulation of gene expression highlight the difficulties in trying to transduce cells in patients. As Kingston and Heagerty point out, the application of gene transfer techniques in complex polygenic disorders is further complicated by uncertainty about which genes to overexpress. We can look to cystic fibrosis, haemophilia, and other diseases with clear molecular targets as barometers of the feasibility of therapeutic overexpression in clinical practice. I for one would be surprised if gene therapy for these conditions becomes established without major advances in the currently available vector systems.


  1. 1-1.
  2. 1-2.
  3. 1-3.
  4. 1-4.
  5. 1-5.