Vector design for expression of O6-methylguanine-DNA methyltransferase in hematopoietic cells

Abstract

Enhancing DNA repair activity of hematopoietic cells by stably integrating gene vectors that express O(6)-methylguanine-DNA-methyltransferase (MGMT) is of major interest for innovative approaches in tumor chemotherapy and for the control of hematopoietic chimerism in the treatment of multiple other acquired or inherited disorders. Crucial determinants of this selection principle are the stringency of treatment with O(6)-alkylating agents and the level of transgenic MGMT expression. Attempts to generate clinically useful MGMT vectors focus on the design of potent expression cassettes, an important component of which is formed by enhancer sequences that are active in primitive as well as more differentiated hematopoietic cells. However, recent studies have revealed that vectors harboring strong enhancer sequences are more likely to induce adverse events related to insertional mutagenesis. Safety-improved vectors that maintain high levels of MGMT expression may be constructed based on the following principles: choice of enhancer-promoter sequences with relatively mild long-distance effects despite a high transcription rate, improved RNA processing (export, stability and translation), and protein design. The need for optimizing MGMT protein design is supported by recent observations suggesting that the P140K mutant of MGMT, developed to be resistant to inhibitors such as O(6)-benzylguanine, may confer a selective disadvantage when expressed at high levels. Here, we provide a review of the literature exploring MGMT expression vectors for bone marrow chemoprotection, and describe experimental evidence suggesting that high expression of MGMT P140K induces a selective disadvantage in the absence of alkylating agents. We conclude that the appropriate design of expression vectors and MGMT protein features will be crucial for the long-term prospects of this promising selection principle.

Figure 1

Principles of gammaretroviral and lentiviral vector design for MGMT expression. The modular composition of the vectors is shown with the respective LTRs (enhancer-promoter (EP), R, U5 regions), splice sites (SD, SA) forming an intron within the expression cassette, rev-responsive element (RRE) and packaging signal (ψ) in the 5′ UTR, internal enhancer-promoter (EP) and MGMT transgene. Optionally a posttranscriptional regulatory element (PRE) is introduced into the 3′ UTR for improved mRNA processing. (A) shows a gammaretroviral LTR-driven vector, (B) - (E) display self-inactivating vectors (SIN, which harbor a 3′ U3 deletion) either of the gammaretroviral (B) or lentiviral origin (C-E). Optionally, additional sequences (e.g. cHS4 insulator) can be incorporated into the 3′ U3 deletion and are present also in the 5′ LTR after reverse transcription. In (D) an intron is installed 3′ of the internal promoter. In (E) two expression cassettes are driven by an internal, bi-directional promoter. For further details see text.

Figure 2

Selective disadvantage of transduced 32D cells with high expression of P140K MGMT. We examined the percentage of MGMT-positive cells in 32D cells, a murine myeloid precursor cell line, for up to 65 days without any exposure of alkylating agents (e.g. BCNU, temozolomide). Analysis was performed using a previously described protocol for intracellular MGMT-staining and FACS detection [21]. A set of 5 gammaretroviral (SF91 derivatives) and lentiviral (RRL derivatives) vectors (see Figure 1 for comparison) was used for transduction of 32D cells. Time post transduction is given on the X-axis, percentage of MGMT-positive cells on the Y-axis. Polyclonal populations with high-expressing P140K vectors experience a loss in MGMT marking (SF91.P140K.p and pRRL.PPT.SF91.P140K.p), whereas cells with control constructs with weaker promoters (RRL.PPT.PGK.P140K.p) or the MGMTwt variant do not show this effect.

Figure 3

MGMT G156A does not induce a selective disadvantage. Using the SF91 vector with or without the PRE (p), we examined MGMT or GFP marking in transduced 32D cells for up to 65 days. Again, populations transduced with vectors expressing high levels of MGMT P140K (SF91.P140K with or without PRE) show a significant drop in marking while the control with EGFP (SF91.GFP.p) or the G156A variant does not. A double mutant (P140K/G156A) shows a prolonged drop in MGMT marking over time. Individual time points were monitored in triplicate.