Lentiviral Vectors

FIELD OF THE INVENTION

The invention relates to lentiviral vectors designed to improve in their efficiency of production, transgene capacity, safety profile and utility in target cells. More specifically, the present invention relates to nucleotide sequences encoding a lentiviral vector genome which comprises any one or more of a modified 3′ LTR; a modified 5′ LTR; a vector intron; at least one cis-acting sequence; and/or an interfering RNA. The invention also relates to a lentiviral vector genome comprising any one or more of the modifications described above. Methods and uses involving such a nucleotide sequence or lentiviral vector genome are also encompassed by the invention.

BACKGROUND TO THE INVENTION

The development and manufacture of viral vectors towards vaccines and human gene therapy over the last several decades is well documented in scientific journals and in patents. The use of engineered viruses to deliver transgenes for therapeutic effect is wide-ranging.

Contemporary gene therapy vectors based on RNA viruses such as γ-retroviruses and lentiviruses (Muhlebach, M. D. et al., 2010, Retroviruses: Molecular Biology, Genomics and Pathogenesis, 13:347-370; Antoniou, M. N., Skipper, K. A. & Anakok, O., 2013, Hum. Gene Ther., 24:363-374), and DNA viruses such as adenovirus (Capasso, C. et al., 2014, Viruses, 6:832-855) and adeno-associated virus (AAV) (Kotterman, M. A. & Schaffer, D. V., 2014, Nat. Rev. Genet., 15:445-451) have shown promise in a growing number of human disease indications. These include ex vivo modification of patient cells for hematological conditions (Morgan, R. A. & Kakarla, S., 2014, Cancer J., 20:145-150; Touzot, F. et al., 2014, Expert Opin. Biol. Ther., 14:789-798), and in vivo treatment of ophthalmic (Balaggan, K. S. & Ali, R. R., 2012, Gene Ther., 19:145-153), cardiovascular (Katz, M. G. et al., 2013, Hum. Gene Ther., 24:914-927), neurodegenerative diseases (Coune, P. G., Schneider, B. L. & Aebischer, P., 2012, Cold Spring Harb. Perspect. Med., 4:a009431) and tumour therapy (Pazarentzos, E. & Mazarakis, N. D., 2014, Adv. Exp. Med Biol., 818:255-280).

As the underlying causes of many genetic diseases are being revealed, it is clear that the delivery of more functionality to the genetic payload (rather than a single gene) within vector genomes is becoming extremely desirable. Thus, there is expectation that transgene cassettes will become more complex, requiring the delivery of more functions, for example in delivering more genes, transgene control (e.g. gene switch systems or inverted transgene expression cassettes) or suicide switches.

The current ‘limits’ of lentiviral vector capacity have not changed significantly over the last 20 years, and remain in the region of ˜7 kb of transgene space when employing standard genome cis-acting sequences such as the typical packaging sequence, rev-response element (RRE) and post-transcriptional regulatory elements (PREs) such as that from the woodchuck hepatitis virus (wPRE). Intrinsically, some aspect of this restriction is defined by the size of the wild type HIV-1 genome of ˜9.5kb from which these vector systems are derived. Generally, the specific titres of lentiviral vectors diminish substantially in proportion to their payload size over-and-above this ‘limit’. Several aspects of lentiviral vectorology are likely to contribute to the limit: [1] steady-state pool of vector genomic RNA (vRNA) in the production cell, [2]efficiency of conversion of vRNA to dsDNA by reverse transcriptase, and [3] efficiency of nuclear import and/or integration into host DNA. The desire to minimize lentiviral vector backbone sequences has recently lead to attempts to alter the arrangement of existing cis-elements (Sertkaya et al., 2021; Vink et al., 2017) as well as the generation of novel genome configurations to minimize RRE and the packaging signal (WO 2021/181108 A1).

As discussed above, inverted transgene expression cassettes may be desirable. The principal problem with retroviral vectors carrying inverted transgene cassettes that are active during vector production is the production of long dsRNA that forms by base pairing between the viral RNA genome (vRNA) and the mRNA encoding the transgene. The presence of dsRNA within the production cell triggers innate dsRNA sensing pathways, such as those involving oligoadenylate synthetase-ribonuclease L (OAS-RNase L), protein kinase R (PKR), and interferon (IFN)/melanoma differentiation-associated protein 5 (MDA-5). One solution to avoid this response is to knock-down or knock-out endogenous PKR in the LV production cell, or over-express protein factors shown to inhibit dsRNA sensing mechanisms, as indeed others have shown is possible (Hu et al. (2018) Gene Ther, 25: 454-472; Maetzig et al. (2010), Gene Ther. 17: 400-411; and Poling et al. (2017), RNA Biol. 14: 1570-1579). However, knock-down/-out of these factors may be laborious or difficult, or it may be impossible to achieve the required reduction/loss in activity, and over-expression of protein factors may alter other aspects of the vector production cell, such as viability/vitality, leading to generally less healthy vector production cells.

Retroviruses typically do not utilize strong polyA sequences because there needs to be a balance of transcriptional activity driven from the 5′ LTR and efficient polyadenylation at the 3′ LTR, despite the LTRs being identical in sequence. U3-deleted LTRs have been shown to have less polyadenylation activity compared to wild-type, non-U3-deleted LTRs (Yang et al. (2007), Retrovirology 4:4), indicating that SIN-LTRs within LVs would be limited in the same fashion.

There are several consequences of weak polyadenylation sites within the LTRs of LVs, such as SIN-LTR-containing LVs. In summary, transcriptional read-out of the vector genome expression cassette and/or transgene expression cassette through the polyA sequence within the 3′ LTR into downstream sequences is not efficiently prevented at either the vRNA stage (i.e. in the vector production cell) or the transgene mRNA transcription stage (i.e. in the transduced cell). In addition, transcriptional read-in from cellular genes through the polyA sequence within the 5′ LTR into the vector genome expression cassette is not efficiently prevented at the transgene mRNA transcription stage (i.e. in the transduced cell). Transcriptional read-out and read-in each have deleterious consequences.

In view of the above, there is an ever-present need in the art for viral vectors with improved safety profiles in administration to patients (for example, in the context of vaccination and gene therapy), and/or for improved viral vectors for larger payloads (whilst maintaining suitable titre and safety profiles) and/or for viral vectors with improved efficiency of production. In particular, viral vectors with improved safety profiles, increased payload capacity and improved efficiency of production are urgently needed.

SUMMARY OF THE INVENTION

The present invention is based on the development of lentiviral vectors (LVs) with improved safety profiles, increased payload capacity and/or improved efficiency of production.

As described further herein below, it is intended that one or more of the aspects of the lentiviral vectors according to the invention may be combined in the same vector. It is also intended that one or more of the aspects of the invention may be combined during the production of the same lentiviral vector. Each of the aspects may be used either alone, for example to achieve a particular effect or improvement, or in combination, for example to achieve one or more particular effects or improvements, as required.

Improved Safety Profiles

In a first aspect, the present inventors surprisingly found that employing modified polyadenylation (polyA) sequences within LV genome expression cassettes results in simplified production of vector genomic RNA for packaging, improved transgene expression and reduced transcriptional read-in and -out (both of the vector genome expression cassette and transgene expression cassette) in transduced cells. The use of the modified polyA sequences of the invention is particularly advantageous for LVs flanked by SIN-LTRs due to the reduced polyadenylation activity in SIN-LTRs. The modified polyA sequences of the invention lead to efficient polyadenylation and thus reduced transcriptional read-in and -out of the LV genome expression cassette. The ability to modify polyadenylation sequences and thereby reduce transcriptional read-in and -out of the LV genome expression cassette offers safety advantages over current LV systems.

A foundational element to the invention is, in effect, re-positioning of a polyadenylation signal (PAS) across the U3/R boundary within the 3′ LTR such that the PAS is copied from the 3′ LTR to the 5′ LTR during integration of LVs. Thus, the R region is embedded within the modified polyA sequence. The resulting efficient polyadenylation at the modified polyadenylation sequence will occur at the vRNA 3′ polyA cleavage site, which will be located at the 3′ end of the embedded R region sequence. Sufficient homology (˜20 nucleotides of homology) between the R regions at both 5′ and 3′ ends of the vRNA is provided to allow for efficient first strand transfer during reverse transcription.

The inventors surprisingly found that this modified polyA sequence configuration can be employed to improve transcription termination at the 3′ LTR, whilst simultaneously ensuring that vRNA cleavage (prior to polyadenylation) allows sufficient R homology with the 5′ R region to retain first strand synthesis. When such modified polyA sequences are used in this context, it is found that the requirement for a back-up heterologous polyA sequence downstream of the vRNA expression cassette is avoided. This permits minimization and/or simplification of LV genome constructs (e.g. plasmids) used during vector production.

Further synthetic versions of the modified R-embedded heterologous polyA sequences of the invention can be made by pairing different USEs inserted upstream of the PAS and embedded R sequence with different GU-rich DSE elements inserted downstream of the embedded R sequence. Whilst the USE-PAS sequence residing within the 3′ U3 region will be copied to the 5′ LTR upon integration, the heterologous GU-rich DSE will not be copied. Therefore, to provide the PAS with an efficient DSE in a close position within the LTRs of the integrated LV genome cassette, the 5′ R region of the vRNA is engineered to contain a GU-rich sequence that functions as a DSE in the recapitulated LTRs. Therefore, after reverse transcription and the LTR-copying process, the USE-PAS sequence residing within the U3 region in both 5′ and 3′ LTRs will be ‘serviced’ by this new DSE (i.e. the DSE will act upon the USE-PAS sequence). Optionally, the DSE-modified R sequence can also be employed at the 3′ LTR as part of a synthetic R-embedded heterologous polyA sequence, functioning as the DSE for the 3′ polyA sequence as well.

Altogether, these improvements are referred to herein as “sequence-upgraded pA LTRs (supA-LTRs)”. The supA-LTRs impart improved transcriptional termination to the transgene cassette in target cells (i.e. reduced transcriptional ‘read-out’) and reduced transcriptional ‘read-in’ from upstream cellular promoters, leading to reduced mobilisation of vRNA backbone sequences, and reduced likelihood of transgene cassette interference. In the preferred combination of sequences, the native HIV-1 PAS can be functionally mutated since transcriptional termination within the supA-LTRs is no longer dependent on any HIV-1 sequences present. This ensures that no premature transcription termination is possible in the LV expression cassettes employing the supA-LTRs.

Accordingly, in one aspect, the invention provides a nucleotide sequence comprising a lentiviral vector genome expression cassette, wherein the 3′ LTR of the lentiviral vector genome comprises a modified polyadenylation sequence, and wherein the modified polyadenylation sequence comprises a polyadenylation signal which is 5′ of the 3′ LTR R region.

Increased Payload Capacity

In a second aspect, the present inventors have generated viral vectors with novel short cis-acting sequences in the 3′ UTR of a transgene expression cassette. They have identified two novel short cis-acting sequences that can be introduced into the 3′ UTR of a transgene expression cassette, either alone, or in combination. These novel short nucleotide sequences (and combinations thereof) can either be used in addition to traditional post-transcriptional regulatory elements (PREs e.g. from woodchuck hepatitis virus; wPRE) to boost transgene expression in target cells or to replace these longer PREs entirely, enabling increased transgene capacity whilst maintaining high levels of transgene expression in target cells.

The inventors have surprisingly found that Cytoplasmic Accumulation Region (CAR) sequences previously identified to function within 5′UTR sequences of heterologous mRNA provide enhanced gene expression when incorporated into the 3′UTR of a viral vector transgene expression cassette. Moreover, when located within the transgene expression cassette 3′ UTR, the initially reported 160 bp CAR sequence (composed of 16x repeats of a 10 bp core sequence) could be further minimized to fewer than 16 repeats without loss of the benefit to transgene expression. Surprisingly, these CAR sequences are shown to enhance the transgene expression from transgene cassettes utilizing introns, as well as boosting expression from cassettes already containing a full length wPRE.

The inventors have also identified a minimal ZCCHC14 protein-binding sequence that can be incorporated into the 3′ UTR of a transgene expression cassette to improve transgene expression. The inventors have shown that these minimal ZCCHC14 protein-binding sequences can be combined with the CAR sequences described herein to further enhance transgene expression.

The novel cis-acting sequences described herein can be used to minimize the size of functional cis-acting sequences of all viral vectors such that payloads can be increased and/or titres of vectors containing larger payloads can be improved, whilst maintaining transgene expression levels in target cells. The invention may therefore be employed [1] within viral vector genomes where ‘cargo’ space is not limiting, such that the novel cis-acting sequences further enhance expression of a transgene cassette containing another 3′UTR element, such as the wPRE, or [2] within viral vector genomes where cargo space is limiting (i.e. at or above or substantially above the packaging ‘limit’ of the viral vector system employed), where the novel cis-acting sequences may be used instead of a larger 3′UTR element, such as the wPRE, thus reducing vector genome size, whilst also imparting an increase to transgene expression in target cells compared to a vector genome lacking any 3′UTR cis-acting element.

Accordingly, in a further aspect, the invention provides a nucleotide sequence comprising a transgene expression cassette wherein the 3′ UTR of the transgene expression cassette comprises at least one cis-acting sequence selected from (a) a cis-acting Cytoplasmic Accumulation Region (CAR) sequence; and/or (b) a cis-acting ZCCHC14 protein-binding sequence.

In a third aspect, the present invention is based on the concept of introducing an intron into the vector genome expression cassette in order to enable reduction of the viral backbone sequence. In this regard, the inventors have surprisingly found that introduction of such an intron facilitates removal of the rev-response element (RRE), which allows for more transgene capacity in the vector.

As the resulting vector genomic RNA (vRNA) packaged into vector virions does not contain the intronic sequence, this so-called ‘Vector-Intron’ (VI) is not counted against available ‘space’ on the vRNA. Thus, more space is available for transgene sequences. This new lentivirus (LV) genome configuration is simple to employ, surprisingly does not require rev or an exogenous vRNA-export factor, and may be a more attractive option in moving away from current LV genomes, since most other aspects of LV genome biology remains the same. Additionally, VI may be of further benefit as it is expected that, since the VI is not present in the final integrated LV genome, the potential for mobilisation of vRNA will be reduced compared to RRE-containing LVs, thereby improving the safety profile of the vector.

As described herein, dsRNA species may be formed during production of viral vectors comprising an inverted transgene expression cassette. The present invention solves this problem by providing LV genome expression cassettes that comprise transgene mRNA self-destabilization or self-decay elements, or transgene mRNA nuclear retention signals, that function to reduce the amount of dsRNA formed when the LV genome expression cassette comprises an inverted transgene expression cassette. The transgene mRNA self-destabilization or self-decay elements, or transgene mRNA nuclear retention signals are located within the VI in the LV genome expression cassettes.

The inventors have previously shown (see WO 2021/160993) that the MSD and cryptic splice donor (crSD) in stem loop 2 (SL2) of the HIV-1 packaging sequence within lentiviral vector genome expression cassettes can be extremely promiscuous, leading to aberrant splicing into transgene sequences and resulting in reduction in production of full length vRNA. Surprisingly, as much as 95% of the detectable cytoplasmic mRNA derived from the external promoter driving vRNA production is spliced depending on internal sequences. For efficient vector production, unspliced packageable vRNA is the most desirable product. In addition, the presence of the MSD in the vector backbone delivered in transduced (patient) cells has been shown by others to be utilised by the splicing machinery, when read-through transcription from upstream cellular promoters occurs (lentiviral vectors target active transcription sites), leading to potential aberrant splice-products with cellular exons. Functional ablation of the MSD and crSD appeared to ablate most of this aberrant splicing.

The same functional ablation of this aberrant splicing may be employed in the present invention in order to avoid unwanted spliced products (e.g. MSD to the VI splice acceptor). Moreover, it is surprisingly found that the ability of the VI to impart full RRE-independence of LV genomes is improved by the MSD mutation.

It has been found previously that potential titre losses associated with such mutations can be recovered by supplying a modified U1 snRNA that targets the packaging sequence of the vRNA (WO 2021/160993). Surprisingly, the present inventors also show that the VI is sufficient to recover LV titres without the need to employ the modified U1 snRNA. Therefore, the invention also encompasses use of the VI to increase titres of MSD-mutated LVs. Surprisingly, therefore, it appears that the VI feature and MSD/crSD mutations are functionally symbiotic in generating RRE-deleted LVs. Moreover, the previous finding that MSD-mutated LVs are less prone to transcriptional read-in from cellular gene transcription at the sites of integration is likely to be synergized with the reduced ability for mobilization of VI LV sequences as a consequence of the VI not being present in the final integrated cassette.

Accordingly, in a further aspect, the invention provides a nucleotide sequence comprising a lentiviral vector genome expression cassette, wherein the lentiviral vector genome expression cassette comprises a transgene expression cassette and a vector intron as described herein.

In one aspect, the present invention provides a nucleotide sequence comprising a lentiviral vector genome expression cassette, wherein:

- i) the major splice donor site in the lentiviral vector genome expression cassette is inactivated;
- ii) the lentiviral vector genome expression cassette does not comprise a rev-response element;
- iii) the lentiviral vector genome expression cassette comprises a transgene expression cassette and a vector intron.

In a further aspect, the present invention provides a nucleotide sequence comprising a lentiviral vector genome expression cassette, wherein:

- i) the major splice donor site in the lentiviral vector genome expression cassette is inactivated;
- ii) the lentiviral vector genome expression cassette does not comprise a rev-response element;
- iii) the lentiviral vector genome expression cassette comprises a transgene expression cassette and a vector intron; and
- iv) when the transgene expression cassette is inverted with respect to the lentiviral vector genome expression cassette, the vector intron is not located between the promoter of the transgene expression cassette and the transgene.

In another aspect, the present invention provides a nucleotide sequence comprising a lentiviral vector genome expression cassette, wherein:

- i) the major splice donor site in the lentiviral vector genome expression cassette is inactivated;
- ii) the lentiviral vector genome expression cassette does not comprise a rev-response element;
- iii) the lentiviral vector genome expression cassette comprises a transgene expression cassette and a vector intron; and
- iv) when the transgene expression cassette is inverted with respect to the lentiviral vector genome expression cassette, the nucleotide sequence comprises a sequence as set forth in any of SEQ ID NOs: SEQ ID NOs: 2, 3, 4, 6, 7, and/or 8, and/or the sequences CAGACA, and/or GTGGAGACT.

In another aspect, the present invention provides a nucleotide sequence comprising a lentiviral vector genome expression cassette, wherein:

- i) the major splice donor site in the lentiviral vector genome expression cassette is inactivated;
- ii) the lentiviral vector genome expression cassette does not comprise a rev-response element; and
- iii) the lentiviral vector genome expression cassette comprises a transgene expression cassette and a vector intron, and
- iv) when the transgene expression cassette is inverted with respect to the lentiviral vector genome expression cassette, the 3′ UTR of the transgene expression cassette comprises the vector intron.

In another aspect, the present invention provides a nucleotide sequence comprising a lentiviral vector genome expression cassette, wherein:

- i) the major splice donor site and cryptic splice donor site adjacent to the 3′ end of the major splice donor site in the lentiviral vector genome expression cassette are inactivated;
- ii) the lentiviral vector genome expression cassette does not comprise a rev-response element;
- iii) the lentiviral vector genome expression cassette comprises a transgene expression cassette and a vector intron.

Improved Efficiency of Production

In a fourth aspect, the present inventors surprisingly found that RNA interference (RNAi) targeting a nucleotide of interest (NOI) can be employed in retroviral vector production cells during production of retroviral vectors comprising the NOI without impeding effective expression of the NOI in target cells, the native pathway of virion assembly and the resulting functionality of the viral vector particles. This is not straightforward because the NOI expression cassette and the vector genome molecule that will be packaged into virions are operably linked. Thus, modification of the NOI expression cassette may have adverse consequences on the ability to produce the vector genome molecule in the cell.

As described above, expression of the transgene protein during retroviral vector production may have unwanted effects on vector virion assembly, vector virion activity, process yields and/or final product quality. Furthermore, the formation of double-stranded (ds) RNA (which typically results from opposed transcription within cells) triggers innate dsRNA sensing pathways within the cell leading to loss of de novo protein synthesis. If this occurs during retroviral vector production (e.g. when the retroviral vector genome comprises an inverted transgene expression cassette), this leads to a loss in expression of vector components, and consequently loss in titre.

The present inventors show that RNAi can be employed in retroviral vector production cells to suppress the expression of the NOI (i.e. transgene) during retroviral vector production in order to minimize unwanted effects of the transgene protein on vector virion assembly, vector virion activity, process yields and/or final product quality. Advantageously, the use of RNAi in retroviral vector production cells also permits the rescue of titres of retroviral vectors harbouring an actively transcribed inverted transgene cassette (wherein the transgene expression cassette is all or in part inverted with respect to the retroviral vector genome expression cassette). The inventors surprisingly found that RNAi can be employed during vector production to minimize/eliminate transgene mRNA but not vector genome RNA (vRNA) required for packaging. Thus, the present invention is particularly advantageous for the improved production of retroviral vectors harbouring an actively transcribed inverted transgene cassette.

Accordingly, the present invention provides a single approach to both mediating transgene repression and rescuing titres of vectors containing actively expressed inverted transgene cassettes by the use of RNAi to target the transgene mRNA during retroviral vector production.

Accordingly, in one aspect, the invention provides a nucleotide sequence encoding a lentiviral vector genome, wherein the 3′ LTR of the lentiviral vector genome comprises a modified polyadenylation sequence, and wherein the modified polyadenylation sequence comprises a polyadenylation signal which is 5′ of the 3′ LTR R region.

In a further aspect, the invention provides a nucleotide sequence encoding a lentiviral vector genome, wherein the lentiviral vector genome comprises a modified 5′ LTR, and wherein the R region of the modified 5′ LTR comprises a polyadenylation downstream enhancer element (DSE).

In a further aspect, the invention provides a nucleotide sequence encoding a lentiviral vector genome, wherein the 3′ LTR of the lentiviral vector genome is a modified 3′ LTR as described herein and the 5′ LTR of the lentiviral vector genome is a modified 5′ LTR as described herein.

In a further aspect, the invention provides a nucleotide sequence comprising a lentiviral vector genome, wherein the lentiviral vector genome comprises a modified 3′ LTR and a modified 5′ LTR, wherein the modified 3′ LTR comprises a modified polyadenylation sequence comprising a polyadenylation signal which is 5′ of the R region within the 3′LTR and wherein the R region within the 3′ LTR comprises a polyadenylation DSE, and wherein the modified 5′ LTR comprises a modified polyadenylation sequence comprising a polyadenylation signal which is 5′ of the R region within the 5′LTR and wherein the R region within the 5′ LTR comprises a polyadenylation DSE.

Improved Safety Profiles, Increased Payload Capacity and/or Improved Efficiency of Production

The present inventors have surprisingly found that all of the above aspects of the invention can be used in combination. In particular, two or more aspects of the invention may be used in combination whilst maintaining suitable titre during lentiviral vector production and high levels of transgene expression in target cells. Advantageously, this provides a lentiviral vector having the improvements associated with each individual aspect of the invention which are utilised, i.e. a lentiviral vector having the corresponding improved properties associated with the relevant aspect of the invention. Thus, the present invention provides lentiviral vectors having improved safety profiles and increased payload capacity, improved safety profiles and improved efficiency of production, increased payload capacity and improved efficiency of production, and improved safety profiles, increased payload capacity and improved efficiency of production.

Accordingly, in one aspect, the invention provides a nucleotide sequence comprising a lentiviral vector genome expression cassette, wherein the 3′ LTR of the lentiviral vector genome comprises a modified polyadenylation sequence as described herein, wherein the lentiviral vector genome expression cassette comprises a transgene expression cassette, and wherein the 3′ UTR of the transgene expression cassette comprises at least one cis-acting sequence as described herein.

In a further aspect, the invention provides a nucleotide sequence comprising a lentiviral vector genome expression cassette, wherein the 3′ LTR of the lentiviral vector genome comprises a modified polyadenylation sequence as described herein, wherein the lentiviral vector genome expression cassette comprises a transgene expression cassette and a vector intron as described herein, and wherein the vector intron comprises one or more transgene mRNA self-destabilization or self-decay element(s) as described herein or one or more transgene mRNA nuclear retention signal(s) as described herein.

In a further aspect, the invention provides a nucleotide sequence comprising a lentiviral vector genome expression cassette, wherein the lentiviral vector genome comprises a modified 5′ LTR as described herein, wherein the lentiviral vector genome expression cassette comprises a transgene expression cassette and a vector intron as described herein, and wherein the vector intron comprises one or more transgene mRNA self-destabilization or self-decay element(s) as described herein or one or more transgene mRNA nuclear retention signal(s) as described herein.

In a further aspect, the invention provides a nucleotide sequence comprising a lentiviral vector genome expression cassette, wherein the lentiviral vector genome expression cassette comprises a transgene expression cassette and a vector intron, wherein the 3′ UTR of the transgene expression cassette comprises at least one cis-acting sequence as described herein, and wherein the vector intron comprises one or more transgene mRNA self-destabilization or self-decay element(s) as described herein or one or more transgene mRNA nuclear retention signal(s) as described herein.

In a further aspect, the invention provides a nucleotide sequence comprising a lentiviral vector genome expression cassette, wherein the lentiviral vector genome expression cassette comprises a transgene expression cassette and a vector intron as described herein, and wherein the vector intron comprises one or more transgene mRNA self-destabilization or self-decay element(s) as described herein or one or more transgene mRNA nuclear retention signal(s) as described herein.

In a further aspect, the invention provides a nucleotide sequence comprising a lentiviral vector genome expression cassette, wherein the 3′ LTR of the lentiviral vector genome comprises a modified polyadenylation sequence as described herein, wherein the lentiviral vector genome expression cassette comprises a transgene expression cassette and a vector intron as described herein, wherein the 3′ UTR of the transgene expression cassette comprises at least one cis-acting sequence as described herein, and wherein the vector intron comprises one or more transgene mRNA self-destabilization or self-decay element(s) as described herein or one or more transgene mRNA nuclear retention signal(s) as described herein.

In a further aspect, the invention provides a nucleotide sequence comprising a lentiviral vector genome expression cassette, wherein the lentiviral vector genome comprises a modified 5′ LTR as described herein, wherein the lentiviral vector genome expression cassette comprises a transgene expression cassette and a vector intron as described herein, wherein the 3′ UTR of the transgene expression cassette comprises at least one cis-acting sequence as described herein, and wherein the vector intron comprises one or more transgene mRNA self-destabilization or self-decay element(s) as described herein or one or more transgene mRNA nuclear retention signal(s) as described herein.

In a further aspect, the invention provides a nucleotide sequence comprising a lentiviral vector genome expression cassette, wherein the 3′ LTR of the lentiviral vector genome comprises a modified polyadenylation sequence as described herein, wherein the lentiviral vector genome comprises a modified 5′ LTR as described herein, wherein the lentiviral vector genome expression cassette comprises a transgene expression cassette and a vector intron as described herein, and wherein the vector intron comprises one or more transgene mRNA self-destabilization or self-decay element(s) as described herein or one or more transgene mRNA nuclear retention signal(s) as described herein.

In a further aspect, the invention provides a nucleotide sequence comprising a lentiviral vector genome expression cassette, wherein the 3′ LTR of the lentiviral vector genome comprises a modified polyadenylation sequence as described herein, wherein the lentiviral vector genome comprises a modified 5′ LTR as described herein, wherein the lentiviral vector genome expression cassette comprises a transgene expression cassette and a vector intron as described herein, and wherein the 3′ UTR of the transgene expression cassette comprises at least one cis-acting sequence as described herein.

In a further aspect, the invention provides a nucleotide sequence comprising a lentiviral vector genome expression cassette, wherein the 3′ LTR of the lentiviral vector genome comprises a modified polyadenylation sequence as described herein, wherein the lentiviral vector genome comprises a modified 5′ LTR as described herein, wherein the lentiviral vector genome expression cassette comprises a transgene expression cassette and a vector intron as described herein, wherein the 3′ UTR of the transgene expression cassette comprises at least one cis-acting sequence as described herein, and wherein the vector intron comprises one or more transgene mRNA self-destabilization or self-decay element(s) as described herein or one or more transgene mRNA nuclear retention signal(s) as described herein.

In a further aspect, the invention provides a set of nucleotide sequences comprising nucleotide sequences encoding lentiviral vector components and a nucleotide sequence of the invention.

In a further aspect, the invention provides a set of nucleotide sequences comprising nucleotide sequences encoding lentiviral vector components and a nucleic acid sequence encoding an interfering RNA of the invention.

In a further aspect, the invention provides a set of nucleotide sequences comprising nucleotide sequences encoding lentiviral vector components, a nucleotide sequence of the invention and a nucleic acid sequence encoding an interfering RNA of the invention.

In some embodiments, the set of nucleic acid sequences comprises a first nucleic acid sequence encoding the lentiviral vector genome and at least a second nucleic acid sequence encoding the interfering RNA. Preferably, the first and second nucleic acid sequences are separate nucleic acid sequences. Suitably, the nucleic acid encoding the lentiviral vector genome may not comprise the nucleic acid sequence encoding the interfering RNA.

In some embodiments, the nucleic acid encoding the lentiviral vector genome comprises the nucleic acid sequence encoding the interfering RNA.

In some embodiments, the lentiviral vector components include gag-pol, env, and optionally rev.

In a further aspect, the invention provides a lentiviral vector genome encoded by the nucleotide sequence of the invention.

In a further aspect, the invention provides a lentiviral vector genome as described herein.

In a further aspect, the invention provides an expression cassette comprising a nucleotide sequence of the invention.

Accordingly, in a further aspect, the invention provides an expression cassette encoding a lentiviral vector genome comprising:

- (i) a transgene expression cassette; and
- (ii) a vector intron comprising at least one interfering RNA as described herein.

In a further aspect, the invention provides a viral vector production system comprising a set of nucleotide sequences of the invention.

In a further aspect, the invention provides a cell comprising the nucleotide sequence of the invention, the expression cassette of the invention, the set of nucleotide sequences of the invention or the vector production system of the invention.

In a further aspect, the invention provides a cell for producing lentiviral vectors comprising:

- (a)
  - (i) nucleotide sequences encoding vector components including gag-pol and env, and optionally rev, and the nucleotide sequence of the invention, or the expression cassette of the invention, or the set of nucleotide sequences of the invention; or
  - (ii) the viral vector production system of the invention; and
- (b) optionally, a nucleotide sequence encoding a modified U1 snRNA and/or optionally a nucleotide sequence encoding TRAP.

In a further aspect, the invention provides a method for producing a lentiviral vector, comprising the steps of:

- (a) introducing:
  - (i) nucleotide sequences encoding vector components including gag-pol and env, and optionally rev, and the nucleotide sequence of the invention, or the expression cassette of the invention, or the set of nucleic acid sequences of the invention; or
  - (ii) the viral vector production system of the invention,
  - into a cell; and
- (b) optionally selecting for a cell that comprises nucleotide sequences encoding vector components and the RNA genome of the lentiviral vector; and
- (c) culturing the cell under conditions suitable for the production of the lentiviral vector.

In a further aspect, the invention provides a lentiviral vector produced by the method of the invention.

In a further aspect, the invention provides the use of the nucleotide sequence of the invention, the expression cassette of the invention, the set of nucleotide sequences of the invention, the viral vector production system of the invention, or the cell of the invention, for producing a lentiviral vector.

In a further aspect, the invention provides a lentiviral vector comprising the lentiviral vector genome as described herein.

DESCRIPTION OF THE FIGURES

FIG. 1. The general nucleotide domains comprising typical polyadenylation sequences and the protein complexes that interrogate them, leading to cleavage and addition of polyA tails to mRNA. The Cleavage and Polyadenylation Specificity Factor (CPSF) protein complex binds to the poly(A) signal (PAS; AAUAAA); it contains CPSF1-4 and the associated factors FIP1L and Symplekin (SYMPK). The CPSF3 subunit is the endonuclease acting at the cleavage site. The Cleavage Factor 1 complex (CFIm) recognizes the upstream element (USE); it is composed of NUDT21, CPSF6, and CPSF7. The CSTF complex recognizes the GU- or U-rich downstream element (DSE). CPSF, CSTF, SYMPK, and CFIm interact at the protein level, stabilizing the RNA binding, thus promoting correct cleavage (typically 15-30nt downstream of the PAS, often at ‘CA’ dinucleotides), and PAS recognition and recruitment of the poly(A) polymerase (PAPOLA or PAPOLG). The nuclear poly(A) binding protein (PABPN1) interacts with CFIm and PAPOLA and contributes to the efficiency of polyadenylation.

FIG. 2. The configuration of LTRs (top panel) and resolution of LTRs following reverse transcription (bottom panel) within lentiviral vectors harbouring standard or novel SIN LTRs of the invention (supA-LTRs). Typical 3^rdgeneration lentiviral vector genomic RNA (vRNA) is generated in production cells by transcription from a powerful promoter; the 3′LTR is typically deleted within the U3 promoter region (A) so that no transcription can occur from the LTRs in transduced cells. No ‘internal’ sequences are shown for clarity. [A] Standard lentiviral vector LTRs. Apart from the SIN modification, standard lentiviral vector LTRs are generally not modified further from wild type HIV-1. In the course of reverse transcription, the denoted sequences are copied resulting in two identical LTRs flanking the transgene cassette in the integrated cassette; at the 5′LTR, the sequence between the first nucleotide of R region (dotted lollipop) to the primer binding site (pbs) is copied to the 3′LTR, and at the 3′LTR, the sequence between the polypurine tract (ppu) and the first nucleotide of R region (dotted lollipop) is copied to the 5′LTR. The result of this that the 5′pA site is effectively copied to the 3′LTR. The HIV-1 polyA site is not as strong as some cellular polyAs; however, the deletion of U3 sequences has resulted in removal of polyA USE sequences, making the HIV-1 pA site even weaker. Since lentiviral vectors integrate into transcriptionally active gene regions, the consequence of harbouring a weak pA within the 5′LTR is the potential for transcription read-through (‘read-in’) into the integrated cassette. The consequence of harbouring a weak pA within the 3′LTR is transcription read-through (‘read-out’) into the downstream chromatin. [B] Lentiviral vector LTRs comprising the sequence upgraded polyAs of the invention. To generate ‘sequence upgraded’ pAs (‘supA’) within the flanking LTRs in order to make lentiviral vectors transcriptionally ‘stealthy’, the present invention shows that the 3′LTR can be modified in that a minimal ‘R’ sequence is inserted between the 3′pA and the polyA cleavage site; this modification can be done within a strong heterologous polyA (e.g. SV40 pA) or with a synthetic polyA, and therefore the the wild type R-U5 region can be entirely deleted from the 3′LTR. The consequence of this is that the 3′pA site is effectively repositioned upstream of the first nucleotide of R region (dotted lollipop)—within the U3 or AU3/SIN region—and copied from the 3′LTR to the 5′LTR during reverse-transcription. The use of an USE within the U3 or ΔU3/SIN region may optionally be included to enhance polyadenylation efficiency. In order to associate the new pA site within the ΔU3/SIN region with a DSE (GU-rich box), the 5′LTR is additionally modified to include a new GU-rich box within the first (TAR) loop of the vRNA. This new DSE/GU box is copied to the 3′LTR during reverse-transcription, resulting in stronger pAs within the flanking LTRs, leading to reduced transcription read-in or read-out of the lentiviral vector genome expression cassette. Optionally, the endogenous 5′pA site within the R region of the 5′LTR may be functionally mutated, since it no longer required.

FIG. 3. Modifications to the 3′(SIN)LTR in supA-LTR genomes and the importance of the minimal ‘R’ region. [A] Typical 3^rdgeneration lentiviral vector genomic RNA (vRNA) is generated in production cells by transcription from a powerful promoter; the vRNA comprises from the first nucleotide of 5′ R region (dotted lollipop) to the end of the 3′ R region, down to the polyA cleavage site (and includes a polyA tail added at the cleaved 3′terminus). The vRNA recruits an endognous tRNA to the primer binding site (pbs), which is used as a primer for cDNA synthesis as the first step of reverse-transcription. The RNA sequence of the resulting DNA:RNA hybrid product is degraded by the RNAseH domain of the RT enzyme, allowing 1st strand transfer of the ‘free’ cDNA to the 3′end of the vRNA, where the cDNA anneals to the complementary 3′R region. The remainder of the negative strand cDNA synthesis now proceeds. [B] An example of ‘embedding’ a minimal sequence of the R region between a heterologous pA signal (AATAAA) and a polyA cleavage site within the 3′(SIN)LTR to give a ‘supA-LTR’ in accordance with the invention. Given the approximate distance tollerances of polyA sequences (USE, pA signal, cleavage site, DSE/GU box) reported to enable highly efficient polyadenylation, prior to the invention it was not known if sufficient ‘R’ sequence could be inserted between the pA signal and the cleavage site to allow both efficient polyadenylation and first strand transfer during the RT step.

FIG. 4. LTR reporter constructs used to assess polyadenylation efficiency during testing of novel supA-LTRs. A simple GFP reporter construct was modified at the 3′end to include test LTRs containing polyA sequences, and an internal ribosomal entry site (IRES) and luciferase reporter ORF inserted downstream of this (top panel). Therefore, the level of luciferase expression detected in cells transfected with the reporter constructs inversely correlated with the strength of the test polyA sequences. The general structure of the LTR tested reflected the positions of the heterologous pA signal placed upstream of the R sequence (the first nucleotide of R represented by the dotted lollipop) and downstream of the SIN-U3 sequence (Δ). Other heterologous polyA sequences flanked the pA signal and R region i.e. USE and GU-rich DSE (middle panel). Example ‘R variants’ are shown, which were tested to identify the different lengths of R that could be inserted whilst maintaining polyA activity (bottom panel); R.1-20 comprised the first 20 nucleotides of the R region, R.1-60 comprised the first 60 nucleotides of the R region, and R.1-20c was generated with downstream complementary so that a stable hairpin could be generated. The arrows indicate the likely/possible polyA cleavage region.

FIG. 5. Consequences of partnering the a 3′ supA-LTR modification with either a standard 5′LTR or a 5′ supA-LTR modification. [A] The schematic shows the sequence order of a standard lentiviral vector 5′LTR in production (i.e. unmodified, wild type HIV-1 RU5 sequence) showing the first nucleotide of the R region (dotted lollipop), loop 1 (the TAR stem loop), loop 2 (the polyA stem loop), the polyA cleavage site (CACA) and GU rich DSE (top panel). If a lentiviral vector harbouring this standard 5′LTR was partnered with the modified 3′LTR of the invention (modified 3′ LTR not shown in the to panel), the resolution of LTRs after reverse transcription is as shown (post-transduction); i.e. this configuration exists at both ends of the integrated cassette (bottom panel). The repositioned pA signal (now upstream of R) of the supA-SIN/U3 region is over 100 nucleotides upstream of the endogenous GU-rich DSE in the native R-U5 sequence. [B] In order to associate a DSE with the repositioned 3′ pA signal, the 5′LTR can be modified such that a new cleavage site and GU-rich DSE is engineered into loop 1, so that the stem loop structure of loop 1 is maintained (top panel).,Therefore, after LTR resolution by reverse transcription a strong polyA sequence is generated with (optional) USE, pA signal, cleavage site and GU-rich DSE all in optimal positioning relative to each other (bottom panel). Consequently, the native 5′pA signal within loop 2 can be functionally mutated.

FIG. 6. A schematic to show the functional domains of RNA within the 5′UTR of the wild type HIV-1 genome. [i] A simplified view of the HIV-1 5′UTR indicating the location of the GU-rich DSE embedded between the R and U5 region (primer binding site [arrow] and core packaging region shown). [ii] A more detailed description of the blocks of functional sequences; trans-activation response (TAR) element to which tat binds, enhancing transcription elongation; the polyA region [approximate pA cleavage site denoted]; tRNA-like element (TLE), which loads the tRNA primer onto the primer binding site (PBS); stem-loop 1 (SL1), contains the dimerisation initiation sequence (DIS)); stem-loop 2 (SL2), contains the major splice donor (MSD), and stem-loop 3 (SL3) contains the ‘GxG’ nucleocapsid binding sequence (ψ)—the AUG being the start codon of Gag/Gagpol. [iii] The generalised structure of the entire 5′UTR that approximates the ‘dimerised’ conformation.

FIG. 7. A schematic to show the general features of the improved 5′ and 3′ LTRs of the invention as encoded within a lentiviral vector genome expression cassette. The position of the promoter (Pro) is shown at the 5′LTR, with the transcription start site (TSS) at nucleotide 1 of the R region as indicated by the dotted lollipop. The 5′ R region is modified to include a new GU-rich box (the DSE) downstream of the polyA cleavage region. In this non-limiting example, the region of R comprising 1-20nt is the hashed box region comprising the sequence between the TSS and the cleavage region (this same sequence is embedded in the heterologous pA sequence at the 3′LTR). The loop 2 region contains the native pA signal, which may be optionally mutated/deleted. At the 3′LTR, the self-inactivating (SIN) modification to the U3 region is shown downstream of the 3′polypurine (ppu) tract. Downstream of the SIN-U3 is positioned a heterologous pA sequence (i.e. the native R-U5 sequence is deleted), wherein the 1-20 nucleotide R region sequence—identical to the 1-20 nucleotides of R at the 5′ end—is positioned between the heterologous pA signal and the cleavage region/GU-rich box (DSE) zone. Optionally, an upstream pA enhancer element may be placed between the SIN-U3 and the heterologous pA signal. In effect, relative to standard LV 3′LTRs, the 3′pA signal is repositioned upstream of the first nucleotide of the R region; this means that the 3′pA signal will be copied to the 5′LTR during reverse transcription. The new GU-rich box (DSE) will be copied to the 3′LTR, and will ‘service’ the pA signals of both LTRs. The new LTRs may be optionally used in lentiviral vector harbouring mutations within the major/cryptic splice donor region (MSD/crSD).

FIG. 8. Testing three R region lengths embedded within the SV40 late polyA sequence with regard to polyadenylation efficiency using GFP/Luciferase reporters. The PolyA reporter plasmid described in FIG. 4 was used to test polyadenylation efficiency of ‘R-embedded’ SV40 polyA sequences. HEK293T cells were transfected with EF1a-driven GFP-wPRE-SINLTR-IRES-Gluc containing plasmids, wherein each SINLTR sequence harboured R variants ‘R.1-20’, ‘R.1-60’, ‘R.1-20c’ or ‘no R’ (i.e. just the SV40 pA), or just a standard RU5 containing its own pA and GU-rich DSE in the U5. [A] GFP expression scores were generated by multiplying percentage GFP-positive by the MFI, and Gluc activity was measured in cell lysates. [B] Gluc activity was divided by GFP expression scores to generate normalised Gluc values, which reflected transcriptional read-through of the SINLTR regions.

FIG. 9. Testing three R region lengths embedded within the SV40 late polyA sequence with regard to polyadenylation efficiency and LV titres using GFP/Luciferase LV genome reporters. The R variants ‘R.1-20’, ‘R-160, and ‘R.1-20c’, as well as non-R containing variant ‘SV40 (no R)’ and the native HIV-1 pA knock-out RU5 variant, were cloned into an LV genome polyA reporter construct. This construct is capable of producing LV vector and also report on polyA activity by Gluc assay. HEK293T cells were transfected with the LV genome reporters described, and with LV packaging components to generate LV-GFP/VSVG vector particles. A pPGK-DsRedX plasmid was spiked in to all transfection mixes to measure transfection efficiency. A ‘normal’ LV vector genome was used as a control (lacking the IRES-Gluc-SV40 pA sequence). [A] Post-production cells were measured for GFP and DsRedX expression by flow cytometry and cell lysates were measured for Gluc activity; Gluc activity was normalised by DsRedX expression. [B] Clarified crude LV harvests were titrated on HEK293T cells by flow cytometry to generated TU/mL titres, which were then normalised setting 100% at the ‘RU5’ variant (harbouring a standard 3′LTR).

FIG. 10. Predicted RNA structure of stem loops (SL) found at the 5′end of retroviruses compared to those in engineered lentiviral vector genomes tested in the invention. HIV-1 genomic RNA contains a stem loop at the 5′ terminus called the trans-activation response (TAR) element to which binds tat. Other retroviruses also harbour stem loop structures at their 5′ terminus; RSV and MMTV contain a SL harbouring the GU-rich DSE that ‘services’ a polyA signal encoded between the U3 TATA box and transcription start site. Hybrid TAR/SLs were designed to incorporate RSV or MMTV or a beta-globin polyA based GU-rich DSEs into the HIV-1 TAR stem; the first 18-20 nucleotides of HIV-1 R were retained in order to maintain sequence that may be important for transcription initiation, as well as ensuring at least 18 nucleotides for homology driven first strand transfer. The hybrid structures are drawn based on the ‘1G’ sequence, which is thought to be the primary packaged form of genomic RNA (see Table 1). The GU-rich DSE-modified TAR loop position is depicted in context to the secondary structure of the LV packaging signal (Psi).

FIG. 11. PolyA reporter constructs to assess polyadenylation of different LTR variants in different expression contexts. The schematic displays subtly different polyA reporter constructs used in the study, which were used to assess polyA activity in different settings. The original polyA reporter (shown in its entirety) tests the 3′ supA-(SIN)LTR variants/configurations reflecting polyadenylation at the 3′end of the LV genome cassette in production cells. The ‘R-embedded’ heterologous polyA sequence is shown with a USE and GU-rich DSE. Above this indicates the alternative LTR variants/configurations to reflect either 5′ or 3′ LTRs after reverse transcription and integration into target cells, depending on whether a USE within the SIN-U3 region [i] is included (e.g. {circle around (2)}) or whether loop 1 of the 5′R region of the LV genome (copied to 3′LTR) has been modified to include a new GU-rich DSE [ii], which will ‘service’ the new pA signal upstream of the R region (e.g. {circle around (1)}); the native HIV-1 pA may be optionally deleted from variant LTRs. The 5′LTR reporter constructs also contained extended sequence from the LV genome into the packaging region, including the major splice donor region (SDs). Thus, the 5′LTR reporters modelled transcription ‘read-in’ from upstream cellular promoters and the 3′LTR reporters modelled transcription ‘read-out’ from the transene cassette.

FIG. 12. PolyA reporter constructs to assess polyadenylation of different SIN-LTR variants containing different R region sequences from different retroviruses. The polyA reporter cassette (see FIG. 11) was tested with SIN-LTR variants containing functional sequences as denoted in the grid to the left, modelling either transcriptional ‘read-in’ (5′SIN-LTR) from a cellular promoter or ‘read-out’ (3′SIN-LTR) from the transgene promoter in the context of an integrated cassette. The variants were compared to the standard SIN-LTR (STD SIN-LTR), which just has the deletion in the U3 followed by the native R-U5, comprising the TAR/SL1@ and native polyA signal (grey lollypop)/GU-rich DSE (R-U5). The variants were the R-embedded SV40 polyA sequence, wherein the sequence between the pA signal (PAS; black lollypop)) and GU-rich DSE was replaced with [1] nts 1-20 of the HIV-1 R region (TAR/SL1) or [2] nts 1-45 of RSV R region (containing its own GU-rich DSE) or [3] nts 1-44 of MMTV R region (containing its own GU-rich DSE). The latter two variants were also tested with or without downstream HIV pA signal mutation. The position of the transcription start site (TSS; defines U3-RU5 boundary) is shown by the dotted lollypop. Relative read-through activity was measured in normalised luciferase units (arbitrary units).

FIG. 13. Relative LV titres of vector genomes harbouring GU-rich DSE modified 5′ TAR/SL1 sequences paired with the 3′supA-LTR. Having previously shown that the minimal nt1-20 R-embedded SV40 polyA sequence could be used to generate high titre LV, some of the GU-rich DSE modified hybrid TAR/SL1 variants based on RSV or MMTV R region (see FIG. 10, Table 1) were cloned into the 5′end of the LV genome to assess the impact of both 5′ and 3′ LTR changes on LV titres. LVs encoded GFP were produced in adherent HEK293T cells and titrated by flow cytometry. Titres were normalised to a standard LV vector control (with unmodified 5′ TAR/SL1 and SIN-LTR) and plotted on a log-10 scale.

FIG. 14. Assessment of transcriptional read-in into integrated supA-LTR LV variants of the invention. The supA-LTR variant LVs produced and titrated in FIG. 13, were used to transduce HEK293T cells at matched MOIs, cells passaged for 10 days to ensure unintegrated cDNA was diluted out, and transcription read-in from cellular gene promoters upstream of the 5′SIN/supA-LTR measured by RT-qPCR using primers/probe binding in the packaging region. RNA signal was displayed relative to the standard LV either as total ‘mobilised RNA signal’ or normalised to integrated copy-number.

FIG. 15. PolyA reporter constructs to assess polyadenylation of different supA-LTR variants containing different TAR/SL1-GU/DSE hybrids. The polyA reporter cassette (see FIG. 11) was tested with supA-LTR variants containing functional sequences as denoted in the grid to the left, modelling either transcriptional ‘read-in’ (5′ supA-LTR) from a cellular promoter or ‘read-out’ (3′ supA-LTR) from the transgene promoter in the context of an integrated cassette. The variants were compared to the standard SIN-LTR (STD SIN-LTR), which just has the deletion in the U3 followed by the native R-U5, comprising the TAR/SL1® and native polyA signal (grey lollypop)/GU-rich DSE (R-U5). The variants were the R-embedded SV40 polyA sequence (SV40-R.1-20 3′LTR), wherein the sequence between the pA signal (PAS; black lollypop)) and GU-rich DSE was replaced with the stated TAR/SL1-GU/DSE variants (FIG. 10, Table 1). The variants were also tested with or without downstream HIV pA signal mutation. The position of the transcription start site (TSS; defines U3-RU5 boundary) is shown by the dotted lollypop. Relative read-through activity was measured in normalised luciferase units (arbitrary units).

FIG. 16. PolyA reporter constructs to assess relative contribution of polyadenylation activity by the different modifications introduced in to supA-LTR. The polyA reporter cassette (see FIG. 11) was tested with supA-LTR variants containing functional sequences as denoted in the grid to the left, modelling either transcriptional ‘read-in’ (5′ supA-LTR) from a cellular promoter or ‘read-out’ (3′ supA-LTR) from the transgene promoter in the context of an integrated cassette. The variants were compared to the standard SIN-LTR (STD SIN-LTR), which just has the deletion in the U3 followed by the native R-U5, comprising the TAR/SL1® and native polyA signal (grey lollypop)/GU-rich DSE (R-U5). Variants containing just the 3′ modifications (mod SIN only) were compared to the full recapitulated supA-LTR (mod SIN+mod 5′R), containing the 1GR-GU2 variant. These were also tested with or without downstream HIV pA signal mutation. The position of the transcription start site (TSS; defines U3-RU5 boundary) is shown by the dotted lollypop. Relative read-through activity was measured in normalised luciferase units (arbitrary units).

FIG. 17. A detailed view of the modified 5′ R region and 3′ ‘R-embedded heterologous polyadenylation sequence in the modified 3′SIN-LTR as part of the supA-LTR configuration. The HIV-1 R region from nucleotide 1 to 59 is shown, indicating the three alternative TSSs at the first three nucleotides. Underlined sequence indicates the nucleotides base-paired in stem loop 1 (the TAR loop) and the light grey sequence indicates the loop. The modified 5′ R region exemplified in the invention retains 18-20 nucleotides of the first HIV-1 R 1-20 nucleotides, and introduces additional ‘CA’ motifs downstream to ‘offer’ cleavage sites. The GU-rich DSE of variant ‘GU2’ (see Table 1) is shown as boxed sequence. Also shown as part of the LV expression cassette is the general structure/sequence of the 3′ supA-LTR containing the SIN-U3, optional USE, the PAS (italic) position upstream of the retained ˜20 nucleotides of R.1-20, and a DSE. Since typical LV expression cassettes retain the 3×Gs at the TSS, then expression of 3G, 2G and 1G vRNA can occur in production. Whilst it has been shown for HIV-1 that 1G vRNA preferentially dimerises and is most efficiently packaged of the three variants, the invention allows for all three vector vRNA species to be potential substrates for packaging and subsequent reverse transcription (RT) steps, by ensuring that the 3×Gs are retained downstream of the PAS. However, the invention also discloses a novel promoter that enables just 1G vRNA to be transcribed, and in this case the first 18 nucleotides (i.e. nucleotides 3-to-20 of HIV-1) of the 5′ R region can be inserted directly after the PAS. In this specific case, the two additional Gs (vertical arrows) downstream of the PAS would not necessarily be required. Thus, during 1^ststrand transfer of the new minus strand ssDNA, the 18-20 nucleotides of homology between 5′ and 3′ R regions results in complementarity sufficient to allow annealing between ssDNA and vRNA, and plus strand synthesis initiation with no mismatches at the primer-extention point. The graphic provides further clarity of how the new DSE (boxed) encoded in the 5′ R region is positioned and copied into DNA as a consequence of the RT step, resulting in the USE-PAS-clv-DSE sequences—all within desirable position within respect to each other in both the 5′ and 3′ supA-LTRs.

FIG. 18. Development of a novel CMV/RSV hybrid promoter that generates ‘1G’ LV genomic vRNA. [A] A comparison of the core promoter and transcriptional start site (TSS) sequences for wild type HIV-1 (top sequence), wild type RSV (bottom sequence), and the novel variants engineered as part of the invention. The core TATA box is shown with immediate upstream and downstream flanking sequence. The presence of the 3×G or 1×G at the 5′ terminus of the LV genome vRNA is indicated in each case. Italised sequence is from HIV-1 and bold sequence is from RSV. All other sequence is from the CMV promoter, with spaces indicated by dashes. Variants ‘CMV-RSV1-G’ and ‘CMV-RSV2-1G’ are two hybrid promoters comprising mostly the CMV promoter but with sequence exchanged for the RSV promoter in the core region where indicated. Variant ‘RSV-3G’ is an expression cassette driven by the full RSV promoter i.e. has the standard ‘3G’ 5′ R region of the LV genome. Variant ‘RSV-1G[RSV SL1]’ is also driven by the full length RSV promoter but the RSV R region replaces the HIV-1 5′ R region; this was cloned into an alternative 3′supA-LTR cassette habouring an ‘R-embedded’ SV40 late polyA sequence where the RSV R.1-44 was embedded instead of R from HIV-1 so that 1^ststrand transfer could occur. All other variants were cloned into an LV genome expression cassette with the 3′supA-LTR containing the HIV-1 R.1-20 embedded SV40 late polyA. *Note that all these variant genome cassettes also contained the ‘IRES-Luc’ reporter after the 3′SIN-LTR region and so were longer constructs compared to the standard, unmodified control (black bar). Therefore, relative LV titres are compared to the ‘CMV-3G’ variant, which contained the standard CMV promoter-LTR configuration. [B]Relative LV titres compared to the ‘CMV-3G’ variant.

FIG. 19. Further optimisation of the supA-LTR LV genome expression cassette. [A] A schematic indicating the type/positions of modifications introducing to a standard SIN-LTR LV genome expression cassette to generate a supA-LTR LV genome expression cassette. A typical SIN-LTR encoding LV genome cassette is shown with [1] CMV promoter driving expression of an vRNA genome from a ‘3G’ TSS, standard 5′/3′ RU5 sequences and a ‘back-up’ polyadenylation sequence (SV40 late polyA shown). A series of three independent experiments was performed wherein LVs were produced using genome expression cassettes employing some or all of the supA-LTR features. The novel CMV-RSV2-1G hybrid promoter is employed to generate a ‘1G’ TSS, the 5′ R region modification with GU-rich sequence (1GR-GU2 and 1GR-GU5 variants were tested), the 3′supA-LTR (R.1-20 embedded SV40 late polyA) with or without the USE, and optionally the back-up polyA sequence. [B] The vector titres produced in each of the three separate experiments (STD SIN-LTR in black bars) from using LV genome expression cassettes with the stated variant features (log 10 values plotted on a linear scale).

FIG. 20. The supA-LTR LV genomes enable mutation of the native 5′LTR PAS site and can be paired with ‘MSD2-KO’ LV genomes. The configurations of 5′ and 3′ LTRs are indicated with or without different elements of the invention; all constructs were driven by the CMV promoter (with the ‘3G’ TSS) and the ‘back-up’ polyadenylation sequence was absent except for the standard control (had SV40 late polyA downstream of the 3′SIN-LTR—not shown). The ‘1GR-GU2’ DSE element was positioned in at the 5′ position, and when also optionally used as the embedded ‘R’ sequence in the 3′ supA-LTR sequence the SV40 late polyA GU-rich DSE was deleted to assess the ability of the ‘1GR-GU2’ DSE to functionally replace it. Deletions are indicated by a white X/black box. All combinations of elements were also evaluated in ‘MSD-2KO’ LV genomes, wherein aberrant splicing from the packaging region is eliminated; such genomes produce lower titres but are recovered by co-expression of a modified U1 snRNA (256U1).

FIG. 21. Overview of ‘supA-2pA-LTRs’ employed within LV genomes. The schematic describes how additional polyadenylation sequences can be inserted within the 3′SIN region in the anti-sense orientation, between the ΔU3 region and the R-embedded heterologous polyA sequence, such that it is copied to the 5′SIN-LTR after transcription.

FIG. 22. Defining a minimum R region sequence embedded within a heterologous polyadenylation sequence in the 3′ supA-LTR: effects on polyadenylation/transcriptional read-through. The heterologous SV40 late polyadenylation signal was positioned downstream of the 3′ppt/ΔU3 region within an EF1a-GFP reporter cassette, and upstream of an IRES-luciferase reporter sequence. R region sequences composed of up to 20 nucleotides of HIV-1 R region were inserted between the heterologous PAS and the cleavage site/GU-rich DSE element. The ‘3G-R20’ sequence is the same as ‘R.1-20’ referred to elsewhere in the invention. Truncated variants were generated based on including either the 3×Gs or a single G (‘1G) immediately downstream of the heterologous PAS (modelling use of genomes with ‘3G’ or ‘1G’ vRNAs). Sequences of the embedded R sequences are displayed. Read-through data are normalised luciferase activities, and displayed relative to the 3G-R20/R.1-20 control.

FIG. 23. Defining a minimum R region sequence embedded within a heterologous polyadenylation sequence in the 3′ supA-LTR: effects on LV titres. The configurations of 5′ and 3′ LTRs are indicated with or without different elements of the invention; all constructs were driven by the CMV promoter (with the ‘3G’ TSS) and the ‘back-up’ polyadenylation sequence was absent except for the standard control (had SV40 late polyA downstream of the 3′SIN-LTR—not shown). All 5′ LTRs were mutated in the native HIV-1 5′ PAS and an internal EFS-GFP-wPRE cassette (not shown) but retained the major splice donor. The 5′ R region was either the wild type/standard TAR/SL1 or the modified 5′ R SL1 comprised the 3GR-GU2 variant. The heterologous SV40 late polyadenylation signal was positioned downstream of the 3′ppt/ΔU3 region. R region sequences composed of up to 20 nucleotides of HIV-1 R region were inserted between the heterologous PAS and the cleavage site/GU-rich DSE element. The ‘3G-R20’ sequence is the same as ‘R.1-20’ referred to elsewhere in the invention. Truncated variants were generated based on including either the 3×Gs or a single G (‘1G) immediately downstream of the heterologous PAS (modelling use of genomes with ‘3G’ or ‘1G’ vRNAs). Sequences of the embedded R sequences are displayed in FIG. 22. Vector supernatants were titrated on adherent HEK293T cells, followed by flow cytometry, and data plotted on a log 10 scale.

FIG. 24. Use of supA-LTRs can increase transgene expression in target cells. The configurations of 5′ and 3′ LTRs of the LV genome expression cassettes used during production—as well as the resulting final SIN-LTR generated in target cells—are indicated with or without different elements of the invention. All constructs were driven by the CMV promoter (with the ‘3G’ TSS) and the ‘back-up’ polyadenylation sequence was absent except for the standard control (had SV40 late polyA downstream of the 3′SIN-LTR—not shown). The native HIV-1 5′ PAS and the major splice donor (MSD) were optionally mutated, and an internal EFS-GFP-wPRE cassette (not shown). The 5′ R region was either the wild type/standard TAR/SL1 or the modified 5′ R SL1 comprised the 3GR-GU2 variant. At the 3′ supA(SIN)-LTR, the heterologous SV40 late polyadenylation signal was positioned downstream of the 3′ppt/ΔU3 region. R region sequences composed of 20 nucleotides of HIV-1 R region were inserted between the heterologous PAS and the cleavage site/GU-rich DSE element. The ‘3G-R20’ sequence is the same as ‘R.1-20’ referred to elsewhere in the invention. Alternatively, the 3GR-GU2 sequence was also employed in the 3′ supA-LTR, effectively providing the embedded R sequence, the cleavage site (not shown) and the GU-rich DSE. LVs were produced in suspension (serum-free) HEK293T cells and used to transduce adherent HEK293T cells, followed by analysis of GFP expression by flow cytometry to generate titre values (GFP TU/mL). Adherent HEK293T cells were then transduced at matched MOI, and cell passaged for 10 days, followed by integration assay and flow cytometry. GFP Expression Scores (ES) were generated by multiplying percent positive cells by the median fluorescence intensity (arbitrary units). These were normalised according to average packaging (4f) copy-number per cell, and then compared to the standard LV set to 100%.

FIG. 25. An overview of the positional use of CARe cis-acting elements for use alone or in combination with ZCCHC14 stem loop(s) and/or a PRE within lentiviral vector genomes. The schematic shows the generalized structure of a lentiviral vector genome containing the RRE or a Vector-Intron (i.e. deleted for RRE) and internal transgene expression cassette encoding a gene of interest (GOI); such genomes typically utilize a PRE (such as wPRE) within the transgene 3′ UTR. The PRE may optionally be entirely replaced with minimal CARe sequences alone or in combination with ZCCHC14 stem-loops (ZC′14 SL) up or downstream of the 3′ppt in order to reduce the size of the transgene cassette. For transgene cassettes inverted with respect to the forward directionality of the vector genomic RNA, the same cis-acting element options can be employed in the transgene 3′UTR, except there is no 3′ppt to consider.

FIG. 26. A detailed view of the CARe and ZCCHC14 stem loop sequences and their incorporation into the 3′UTR region of transgene cassettes within viral vectors. A. The consensus sequence for the 10 bp CARe core sequence (or ‘tile’ referred herein). B. Two non-limiting examples of ZCCHC14 binding stem loops found within HCMV (RNA2.7) and WHV (wPRE). ZCCHC14 recruitment leads to formation of a complex with Tent4, which promotes mixed tailing in polyA tails of mRNAs, stabilizing them. C. The concept of insertion of CARe sequences into the 3′ UTR of a viral vector transgene cassette (DNA at top, RNA shown as curvy line below DNA), optionally together with ZCCHC14 stem loops, taking care in retro/lentiviral vectors not to disrupt 3′ppt or att (‘Δ’[i.e. ΔU3]) integration sequences required for reverse transcription and integration respectively. CARe sequences and optionally ZCCHC14 stem loops can be designed rationally or by library design, and screened empirically in target cells. For example screening can be done by viral vector transduction followed by selection of high-expressing cells (e.g. GFP FACS), followed by RT-PCR and sequencing of target mRNA to identify transcripts containing combinations of the cis-acting elements that lead to greater transgene expression and mRNA steady-state pools. This process can be repeated to enrich the best variants, whilst also optionally including error-prone RT-PCR to fine-tune sequences.

FIG. 27. Production titres in suspension (serum-free) HEK293T cells of lentiviral vectors harbouring different transgene promoters combined with 3′ UTR cis-acting elements. A and B present data from two independent experiments for LV-RRE-EFS-GFP vectors containing different 3′ UTR cis-acting elements: wPRE, ΔwPRE (wPRE deleted), 16×10 bp CARe sequences in sense (CARe.16t) or antisense (CARe.inv16t) and/or single copy of the ZCCHC14 stem loop from HCMV RNA2.7 (HCMV.ZSL1). C shows data for output titres of LV-RRE-EF1a-GFP (EF1a contains an intron) and LV-RRE-huPGK-GFP. Titres were measured by transduction of adherent HEK293T cells followed by flow cytometry based assay after 3 days (GFP TU/mL) or qPCR to LV DNA after 10 days (Integrating TU/mL). The data shows that integrating titres of LVs are comparable irrespective of the presence/absence of any of the cis-acting elements but that GFP titres vary, reflecting the expression levels in transduced adherent HEK293T cells. The 16×10 bp CARe tile (only) in the sense orientation provided a boost to LV GFP TU/ml titres lacking the wPRE.

FIG. 28. The 16×10 bp CARe tile boosts transgene expression from lentiviral vectors lacking wPRE in a T-cell line. LV-RRE-EFS-GFP and LV-RRE-EF1a-GFP vector stocks produced in suspension (serum-free) HEK293 Ts were used to transduce Jurkat cells at matched multiplicity of infection (MOI): MOI 1 [A], MOI 0.25 [B] and MOI 0.1 [C]. Transduced cells were analysed by flow cytometry to obtain % GFP-positive values and median fluorescence intensity values (Arbitrary units). D displays data from normalized RT-PCR of extracted mRNA from the transduced cells, where the 100% level is set for each EFS-GFP or EF1a-GFP cassette containing the wPRE in each case. The data show that the 16×10 bp CARe tile (only) in the sense orientation restores transgene expression levels to those observed with wPRE-only, and for EF1a-GFP surprisingly boosts transgene expression levels higher the wPRE-only. The 16×10 bp CARe tile (only) in the sense orientation increase the levels of transgene mRNA above wPRE-only in all conditions.

FIG. 29. Production titres in suspension (serum-free) HEK293T cells of ‘MSD-2KO’/‘U1-dependent’ lentiviral vectors harbouring different transgene promoters combined with 3′ UTR cis-acting elements. LV-RRE-Pro-GFP vectors containing mutations in the SL2 loop of the packaging signal (thus ablating aberrant splicing from this region) were produced+/−256U1, a modified U1 snRNA that binds to the vector genomic RNA to restore titres. Three different transgene promoters (EFS, EF1a, and huPGK) and different 3′ UTR cis-acting elements were employed: wPRE, ΔwPRE (wPRE deleted), and 16×10 bp CARe sequences in sense (CARe.16t) or antisense (CARe.inv16t). Titres were measured by transduction of adherent HEK293T cells followed by flow cytometry based assay after 3 days (GFP TU/mL) or qPCR to LV DNA after 10 days (Integrating TU/mL).

FIG. 30. Production titres in suspension (serum-free) HEK293T cells of ‘MSD-2KO’/‘/ΔRRE’ lentiviral vectors harbouring different 3′ UTR cis-acting elements and transgene expression in target cells. LV-VI(ΔRRE)-EFS-GFP vectors containing mutations in the SL2 loop of the packaging signal (thus ablating aberrant splicing from this region) were produced. Different 3′ UTR cis-acting elements were employed: wPRE, ΔwPRE (wPRE deleted), 16×10 bp CARe sequences in sense (CARe.16t) or antisense (CARe.inv16t), and/or single copy of the ZCCHC14 stem loop from either HCMV RNA2.7 (HCMV.ZSL1) or WHV wPRE (WPRE.ZSI1). A. Titres were measured by transduction of adherent HEK293T cells followed by flow cytometry based assay after 3 days (GFP TU/mL) or qPCR to LV DNA after 10 days (Integrating TU/mL). B. Transgene expression in transduced adherent HEK293T or HEPG2 cells was measured by flow cytometry three days post-transduction at match MOI.

FIG. 31. Transgene expression levels in primary cells transduced with RRE/rev-dependent lentiviral vectors harbouring different 3′ UTR cis-acting elements. LV-RRE-EFS-GFP [A] or LV-RRE-EF1a-GFP [B] vector stocks produced in suspension (serum-free) HEK293 Ts were used to transduce primary cells (92BR) at matched multiplicity of infection (MOI): MOI 2, 1 or 0.5. Different 3′ UTR cis-acting elements were employed: wPRE, wPRE3 (shortened wPRE), ΔwPRE (wPRE deleted), 16×10 bp CARe sequences in sense (CARe.16t) or antisense (CARe.inv16t), and/or single copy of the ZCCHC14 stem loop from either HCMV RNA2.7 (HCMV.ZSL1) or WHV wPRE (WPRE.ZSI1). Transgene expression in transduced adherent 92BR cells was measured by flow cytometry three days post-transduction at match MOI.

FIG. 32. Transgene expression levels in adherent HEK293T cells transduced with RRE/rev-dependent lentiviral vectors harbouring different 3′ UTR cis-acting elements at matched MOIs. LV-RRE-EFS-GFP vector stocks produced in suspension (serum-free) HEK293 Ts were initially titrated on adherent HEK293T cells to generate integrating titres (TU/mL). Vector stocks were used to transduce fresh adherent HEK293T cells at matched multiplicity of infection (MOI): MOI 2, 1 or 0.5. Different 3′ UTR cis-acting elements were employed as ‘stand-alone’ elements: wPRE, 16×10 bp CARe tiles (CARe.16t) or a single copy of the ZCCHC14 stem loop (HCMV.ZSL1), compared to no element (ΔwPRE). Additionally, variants deleted for wPRE but containing a single copy of the ZCCHC14 stem loop were also paired with increasing numbers of CARe tile, from 1× to 20×10 bp copies. Transgene (GFP) expression in transduced adherent HEK293T cells was measured by flow cytometry three days post-transduction and median fluorescence intensities (Arbitrary units) normalised to that achieved with the standard wPRE-containing LV (set to 100%).

FIG. 33: A schematic comparing DNA expression cassettes for standard and Vector-Intron containing LV genomes and the mRNAs transcribed therefrom. The general structure of typical standard 3^rdgeneration LV genomes is shown, containing: a U3-deleted, tat-independent heterologous promoter driving transcription (Pro), the broad packaging sequence from R-U5 to the gag region, the RRE, the central polypurine tract (cppt), an internal transgene expression cassette (Pro-GOI), a post-transcriptional regulatory element (PRE) and a self-inactivating 3′LTR. The SL1 loop of the broad packaging sequence contains the MSD and adjacent crSD. The core packaging motif (4J) is within SL3. The amount of retained gag sequence can vary but is typically between -340 and -690 nts from the primary ATG codon of gag, and includes the p17 instability element (p17-INS). The RRE is typically in the region of 780 bp, and includes the splice acceptor ‘7’ site (sa7) from HIV-1. For standard LV genome cassettes, apart from the main transgene mRNA (assuming the internal promoter is active in production cells), the primary transcript produced and exported to steady-state levels in the cytoplasm by rev was thought to be the full length vRNA. However, the inventors have shown elsewhere that promiscuous or aberrant splicing from the MSD or the crSD in the SL2 loop occurs to (cryptic) splice acceptors downstream of sa7 even in the presence of rev. The amount of spliced product compared to full length vRNA can be 20:1, especially when the transgene cassette contains a strong splice acceptor such as the one present in the EF1a promoter. The novel LV genome of the present invention replaces the RRE entirely with a single intron in order to increase transgene payload, since the intronic sequence will be absent from the full length vRNA. The MSD/crSD mutation ensures that no aberrant splicing from SL2 can occur with the VI splice acceptor. Surprisingly, not only does the act of splicing out of the VI allow vRNA to be stabilized in a rev/RRE-independent manner, it also abrogates the attenuating effect of MSD/crSD mutation on LV titres. Further, it is shown that RRE-deleted VI genomes achieve greatest titres in a rev-independent manner when the MSD/crSD is mutated. The novel LV genome may also incorporate a deletion of the p17-INS, therefore also increasing transgene capacity by a total of ˜1kb.

FIG. 34: Rev/RRE-independent HIV-1 based LVs containing a Vector-Intron are improved by mutation of the major splice donor and cryptic splice donor sites in SL2 of the packaging signal. HIV-1 based LV genomes (with an internal CMV-GFP cassette) were generated containing various combinations of either standard or mutated/deleted cis-acting elements (STD-MSD or MSD-2KO, +RRE, Vector-Intron; see FIG. 33). These genome plasmids were used to produce LV-CMV-GFP vectors in either adherent (Δ) or suspension [serum-free] (B) HEK293T cells in the presence or absence of a rev-expression plasmid. Clarified vector supernatants were titrated on adherent HEK293T cells using flow cytometry, and vector titres plotted on a log 10 scale.

FIG. 35: Analysis of vector cassette-derived RNA in adherent production cells and in resulting vector particles for variant genomes containing a Vector-Intron in combination with other cis-elements/mutations. Total extracted RNA from production cells and vector particles from the adherent cell production run of Vector-Intron (VI_v1.1) genomes described for FIG. 34A was subjected to RT-PCR to assess the species produced by each genome variant (panel B). The DNA, pre-RNA and main splicing products for these four Vector-Intron genome variants is shown schematically in panel A. The splicing of the Vector-Intron is denoted as well as the potential aberrant splicing of the MSD to the VI splice acceptor. The optional presence of the RRE is also denoted, as well as the positions of the PCR primers (grey arrows) used for the RT-PCR analysis (oligo-dT primer was used for the cDNA step).

FIG. 36: A schematic showing the core features of exon-intron-exon sequences important for splicing. The schematic shows a representative single intron (grey block) between two exons (black blocks), indicating the position of key consensus sequences required for splicing-out of the intron, as well as the position of enhancers. The termini of the intron are defined by GT-AG dinucleotides at the 5′ and 3′ ends respectively. The GT dinucleotide is the least variable sequence of the broader splice donor consensus sequence; the consensus sequence is the target of U1 snRNA, which anneals to the donor site early on during exon/intron boundary recognition. The AG dinucleotide is the least variable sequence of the broader splice acceptor consensus sequence, which typically comprises a polypurine tract of ˜12-to-20 nts upstream. The Branch site (consensus=TNCTRAC, wherein “N” means any nucleotide and “R” means A or G) is located 20-to-35 nts upstream of the spliced acceptor site and is the target of U2 snRNA, which anneals to the branch site during the splicing reaction. The Branch site is also the anchor point for the linkage of the 5′ end of the intron to form the lariat structure. The length of the intron can be short or many thousands of nucleotides, and may contain other cis-acting elements, with some partaking in enhancing or regulating splicing efficiency. Intronic splicing enhancers (ISEs) may be located closer to the ends of the intron so as to be in close proximity to the core elements described above. Exonic splicing enhancers (ESEs) may also be located close to the exon-intron junction in order to mediate effects. In the present invention, a number of these functional elements from different organisms were evaluated towards the optimization of the Vector-Intron approach.

FIG. 37: Assessment of initial Vector-Intron variants in HIV-1 based LVs in adherent HEK293T production cells. Genome plasmids harbouring an EFS-GFP transgene cassette but lacking the RRE were constructed to contain the MSD-2KO mutations and six variant Vector-Introns, as per Table 7. These were based on native introns from the EF1a or Ubiquitin (UBC) promoter-introns, or the CAG promoter-intron or the small chimeric intron (Syn) from the pCI series of expression plasmids by Promega. These genome plasmids were used to produce LV-EFS-GFP vectors in adherent HEK293T cells in the absence of a rev-expression plasmid, whereas a standard LV vector was made +/−rev. Clarified vector supernatants were titrated on adherent HEK293T cells using flow cytometry, and vector titres plotted on a log 10 scale.

FIG. 38. Further development of Vector-Intron variants based on a chimeric intron in HIV-1 based LVs in suspension (serum-free) HEK293T production cells. Genome plasmids harbouring an EFS-GFP transgene cassette but lacking the RRE were constructed to contain the MSD-2KO mutations and seven variant Vector-Introns VI_v4.2-4.8, as per Table 7. These were based on the small chimeric intron from the pCI series of expression plasmids by Promega but varied mainly in the presence/type of upstream ESE and/or splice donor sequence. These genome plasmids were used to produce LV-EFS-GFP vectors in suspension (serum-free) HEK293T cells in the absence of a rev-expression plasmid, whereas a standard LV vector was made +/−rev. Clarified vector supernatants were titrated on adherent HEK293T cells using flow cytometry, and vector titres plotted on a log 10 scale.

FIG. 39. Further development of Vector-Intron variants based on the human p-globin intron-2 in HIV-1 based LVs in suspension (serum-free) HEK293T production cells. Genome plasmids harbouring an EFS-GFP transgene cassette but lacking the RRE were constructed to contain the MSD-2KO mutations and two Vector-Introns VI_v5.1/5.2 based on the second (truncated) intron of the human p-globin gene, as per Table 7. These were compared to two of the previous Vector-Intron variants based on the chimeric intron from the pCI series of Promega plasmids (VI_4.2/4.8). These genome plasmids were used to produce LV-EFS-GFP vectors in suspension (serum-free) HEK293T cells in the presence/absence of a rev-expression plasmid, whereas a standard LV vector was made +rev. Clarified vector supernatants were titrated on adherent HEK293T cells using flow cytometry, and vector titres plotted on a log 10 scale.

FIG. 40. Testing of LV genomes with Vector-Introns in combination with different MSD-mutations and p17-INS(gag) deletion. Vector-Intron variants from two series (v4 [pCI] and v5 [hu β-Globin]) were tested in LV genomes in the context of two different MSD-mutation variants (‘MSD-2KO’ and ‘MSD-2KOm5’), and additionally with the p17-INS deleted from the gag region of the packaging sequence. These genome plasmids were used to produce LV-EFS-GFP vectors in suspension (serum-free) HEK293T cells in the absence of a rev-expression plasmid, whereas a standard LV vector was made +/−rev. Clarified vector supernatants were titrated on adherent HEK293T cells using flow cytometry, and titres normalized to the Standard LV-GFP vector prep made with rev.

FIG. 41. Evaluation of titre increase by Prostratin on Vector-Intron LV genomes. Standard or Vector-Intron/MSD-2KO/ΔRREΔp17-INS genome plasmids were used to produce LV-EFS-GFP vectors in suspension (serum-free) HEK293T cells in the absence of a rev-expression plasmid, whereas a standard LV vector was made +/−rev. Vectors were made in the presence or absence of 11 μM Prostratin −20 hours post-transfection (at sodium butyrate induction step). Clarified vector supernatants were titrated on adherent HEK293T cells using flow cytometry, and vector titres plotted on a linear scale.

FIG. 42. Rev-independent production of Vector-Intron LVs containing different transgene promoters. HIV-1 based LVs containing CMV/EFS driven transgene cassettes within either standard or Vector-Intron backbones were produced to high titres in suspension (serum-free) HEK293T cells, in the presence or absence of rev, respectively. Vector titres are plotted on a log 10 scale.

FIG. 43. Utilisation of LV genome cassettes containing inverted transgene cassettes expressed during LV production leads to suppression of de novo vector component protein expression via a cytoplasmic dsRNA sensing mechanism. Standard RRE-containing LV genome plasmids (STD RRE-LV) or Vector-Intron genome plasmids (MSD-2KOm5/ΔRRE/Δp17-INS+VI_v5.5) [both 3G TSS constructs unless indicated] were generated with either forward (Fwd) or inverted (Invert) EFS-GFP transgene cassettes, and used to produce LV harvest supernatants in suspension (serum-free) HEK293T cells. Packaging plasmids (pGagPol and pVSVG) were co-transfected together with or without pRev were indicated. Supernatants were analysed by SDS-PAGE/immunoblotting to VSVG and p24 (capsid). The data indicate that inverted transgene cassettes induce suppression of de novo LV component synthesis, consistent with a cytoplasmic dsRNA sensing mechanism e.g. PKR.

FIG. 44. A schematic showing an example of a Vector-Intron LV genome with inverted transgene cassette. A Vector-Intron LV genome with a reverse facing transgene cassette containing an intron is shown. Since the VI stimulates intron loss only from the vRNA (top strand-copied), the transgene cassette will retain its own intron. Depending on the strength of the transgene cassette promoter, a significant amount of double-stranded RNA may form between the vRNA and the transgene mRNA during LV production. This can potentially lead to a PKR response, cleavage by Dicer or deamination by ADAR; any or all of these mechanisms can contribute to reduced vector titres. This can be avoided by utilizing the unique features of the VI by inserting into it cis-acting element(s) (X) within the 3′UTR of the inverted transgene cassette. Such cis-acting elements are those that would reduce the abundance of only the transgene mRNA e.g. AU-rich [instability] elements (AREs), miRNAs, and/or self-cleaving ribozymes. The action of these reduce the amount of transgene mRNA available for pairing with the complementary vRNA to generate dsRNA. In addition, reduced transgene mRNA (and resultant protein) can be advantageous for LV production. Importantly, the cis-acting element(s) will not be present within the final integrated transgene cassette due to out-splicing of the VI, and therefore transgene mRNA stability in the transduced cell will be efficient.

FIG. 45. A schematic showing an example of a Vector-Intron LV genome with inverted transgene cassette and further details of cis-acting elements within the 3′UTR of the transgene cassette that mediate transgene mRNA degradation. A Vector-Intron LV genome with a reverse facing transgene cassette containing an intron is shown during LV production. The use of ‘functional’ cis-acting elements (‘X’) within the 3′UTR of the transgene cassette—and located within the anti-sense VI sequence—can be used to achieve transgene repression and to avoid dsRNA responses during LV production. Two examples of functional cis-acting sequences are shown. Firstly, one or multiple self-cleaving ribozymes (‘Z’) can be inserted within the anti-sense VI sequence of the 3′UTR, leading to self-cleavage of pre-mRNA, resulting in RNA lacking a polyA tail and degradation in the nucleus. Secondly, one or multiple pre-miRNAs (‘m’) can be inserted within the anti-sense VI sequence of the 3′UTR, leading to pre-miRNA cleavage/processing resulting in cleavage of the pre-mRNA. Importantly, the miRNAs can be targeted to the transgene mRNA so that any mRNA that does locate to the cytoplasm is a target for microRNA-mediated cleavage (the guide strand should be 100% matched to its target). The vRNA will not be targeted by the guide strand. The passenger strand should be mis-matched with regard to the vRNA sequence to avoid cleavage of the vRNA should the passenger strand become a legitimate microRNA effector. Thus, both of these examples can be used to reduce/eliminate transgene mRNA (and dsRNA) only in LV production, since these functional cis-acting elements will be lost from the packaged vRNA due to loss of the VI.

FIG. 46. Use of self-cleaving ribozymes within the 3′UTR of inverted transgene cassettes to enhance LV virion production. LV genome cassettes containing an inverted EF1a-GFP transgene with or without ‘functionalized’ 3′ UTRs were used to produce LVs in suspension (serum-free) HEK293 Ts, and vector proteins components within clarified harvest material analysed (panel A). Levels of vRNA were assessed by RT-PCR in both post-production cells (‘C’) and vector supernatant harvest (‘V’). Expression of an inverted transgene cassette during LV production leads to double-stranded (ds) RNA, since the mRNA will be complementary to the majority of the LV vRNA. dsRNA is likely triggering at least one sensing mechanism during production (e.g. PKR), leading to a substantial reduction in detectable VSVG and p24 (capsid) in harvest material (panel A) and vRNA (panel C)—see ‘Empty’ lanes. Four different self-cleaving ribozymes were tested within the Vector-Intron (VI) region of the 3′UTR of the inverted transgene cassette: Hammerhead ribozyme (HH_RZ), Hepatitis delta virus ribozyme (HDV-AG), and modified Schistosoma mansoni hammerhead ribozymes (T3H38/T3H48) (see panel [B] for schematics and [A, C for data]. Additionally, variants were made harbouring the ‘negative regulator of splicing’ (NRS) element from RSV, or a splice donor site, as these have been shown to impart destabilization effects on mRNA. Other variants included several of these cis-acting elements in the same 3′UTR, with upstream/downstream positions noted as [1] or [2] respectively. (The positions of the forward [f] and reverse [r] primers for RT-PCR analysis is indicated in panel B; other features such as cppt and wPRE are not shown). These elements were cloned into an LV genome containing the VI_5.7 variant, the MSD-2KOm5 modification and deletions in the gag-Psi and RRE (full deletion) regions. Vectors were produced alongside the standard, RRE/rev-dependent LV genome containing the cassette in the forward direction, which produces aberrant splice products (see panel C). The data show that use of self-cleaving ribozymes enables recovery of VSVG, p24 and vRNA with LV supernatants.

FIG. 47. Use of self-cleaving ribozymes within the 3′UTR of inverted transgene cassettes to enhance LV titres. LV harvest supernatants described in Example 20 and FIG. 46 were titrated by integration assay in HEK293T cells. The data demonstrated that titres of VL LV genomes harbouring active inverted transgene cassettes are ˜1000-fold lower than STD RRE-LVs containing the same transgene in the forward (fwd) orientation (see ‘Empty’). However, the use of self-cleaving ribozymes within the 3′ UTR of the inverted transgene cassette enables ˜100-fold recovery in titres. The presence of other cis-elements NSR and a splice donor site had no/minimal effect on titres.

FIG. 48. Use of minimal gag sequences as part of the packaging signal within Vector-Intron LV genomes. The retained gag sequence within the packaging region of Vector-Intron genomes (reduced to 81 in other examples) was further minimized, resulting in constructs harbouring 57, 31, 14 or zero nucleotides of gag. All constructs harboured an ATG>ACG mutation in the primary initiation codon of retained gag sequence. All variants were presented within an MSD-2KOm5 LV genome containing VI_5.5 in place of the RRE. The standard LV genome contained the MSD, RRE and the first 340 nucleotides of gag (including the p17INS). LVs were produced in suspension (serum-free) HEK293T cells by transient transfection, and clarified harvests titrated on adherent HEK293T cells followed by flow cytometry.

FIG. 49. Optimisation of rev-independent production of Vector-Intron LVs: fine-tuning of vector component input levels. A. Clarified standard (STD) or Vector-Intron (VI) LV vector supernatants (duplicate) were immunoblotted using antibodies to VSVG (white) or p24/capsid (black). M=molecular weight markers (kDa). B. A Design-of-Experiment (DoE) multivariate analysis experiment was performed using an MSD-2KOm5/Δp17INS/VI_5.5 LV genome encoding EFS-GFP. The control/centre point set of ratios (genome:gagPol:VSVG of 950:100:70 ng/mL) was the optimized ratio used for standard RRE/rev-dependent LVs (and all previous examples assessing VI LVs). LVs were produced LVs were produced in suspension (serum-free) HEK293T cells by transient transfection, and vector was harvested at the time points (hours) stated post-induction with sodium butyrate. Clarified harvests titrated on adherent HEK293T cells followed by flow cytometry.

FIG. 50. A schematic showing how microRNA targeted against the transgene mRNA of a lentiviral vector (LV) containing an inverted transgene cassette can be used to avoid production of dsRNA, and to reduce transgene expression. The configurations of both forward facing and inverted transgene cassettes with LV genome expression cassettes are indicated, as are the packaged vRNA (ψ) and transgene mRNAs in each case. Use of inverted transgene cassettes within retroviral vectors typically leads to a reduction in vector production due to the generation of long dsRNA; this typically induces dsRNA sensing pathways in the cell (such as PKR-mediated translation suppression), leading to reduction in vector component protein expression. To avoid this, one or more microRNAs can be co-expressed during vector production (e.g. by co-transfection with siRNA or with a microRNA expression cassette), wherein the microRNA targets the transgene mRNA for cleavage. Use of a mis-matched passenger strand can avoid loss of vRNA due to low level loading of the passenger strand as the guide within the RISC. Cleavage (and resultant degradation) of transgene mRNA leads to reduction in transgene protein expression during LV production, which can be advantage in achieving maximal titres and/or product recovery/purity.

FIG. 51. A schematic showing the different microRNA ‘modalities’ that can be adopted in the invention. The transgene-targeting microRNA can be part of a ‘transient’ or ‘stable’ vector process using cell transfection or stable cell lines, respectively. For transient transfection approaches the microRNA can be delivered as siRNA or shRNA, or as a miR expression cassette, where the microRNA is transcribed de novo, for example, from a polymerase-III promoter such as U6 or a tRNA promoter. The miR cassette may be a separate plasmid or alternatively could be inserted within the vector genome plasmid or packaging plasmids. The miR may also be stably integrated into the production cell, which itself may or may not also contain the all or some of the vector components.

FIG. 52. Production of LVs using siRNA to repress transgene expression from forward facing or inverted transgene cassettes. LVs containing an EFS-promoter driven GFP cassette either in the forward (Fwd) or inverted (Invert) orientation were produced in suspension (serum-free) HEK293T cells. Production cells were co-transfected with LV genome and packaging plasmids with or without the stated siRNAs, as well as a DsRed-Xprs reporter plasmid control, and post-production cells analysed by flow cytometry for GFP/DsRed-Xprs expression levels (% positive gate x median fluorescence intensity; Arbitrary units). Clarified vector supernatants were titrated by transduction of adherent HEK293T cells followed by flow cytometry (Titre in TU/mL). The control siRNA was directed to Luciferase (not present), and the DsRed-Xprs reporter was present to assess the impact of dsRNA production on de novo protein synthesis.

FIG. 53. Building a supA-2pA-LTR: insertion of inverted polyadenylation signal and inverted GU-rich DSE within the SINΔU3 region to reduce transcriptional read-in from 3′ end of integrated LVs. A number of variants of SIN-LTR-like composite sequences were generated based upon positioning differing lengths of the SV40 polyadenylation sequence downstream of the SINΔU3 (attΔU3) sequence, and harbouring mutations in different polyA signals (pAm1) present upstream of the RU5 (where indicated) or the native HIV-1 polyA signal (where indicated). The SV40 polyadenylation sequence is bi-directional, with the arrow indicating the direction of the late sequence, which included the late USE and polyA signal but not the late GU-rich DSE. Note that the modified (5′) R—containing the optimal GU-rich DSE—could in principle have been use in place of the TAR to improve polyadenylation as shown elsewhere where but was not in this initial example. In previous examples of supA-LTR sequences the late SV40 polyA sequence (containing the late USE-PAS sequence) includes the two polyA signals of the early SV40 polyA sequence on the bottom strand (i.e. inverted) but the early GU-rich DSE is omitted. Constructs 2-7 model the ‘top-strand’ orientation (i.e. transcriptional read-in from upstream cellular promoters), whereas constructs 8-12 model ‘bottom-strand’ orientation (i.e. transcription read-in from downstream cellular promoters), although the inverted RU5 sequence was not present in these inverted variants in this example. To provide a GU-rich DSE for the early SV40 polyA sequence, the native SV40 early polyA sequence was simply extended to include the native GU-rich DSE (contructs 6, 10, 11) or a variant GU-rich DSE based on the MMTV GU-rich DSE was inserted downstream of the two early PAS's (contructs 7, 12). These sequences were inserted into the luciferase polyA reporter and suspension (serum-free) HEK293T cells transfected, followed by luciferase assay of cell lysates to measure transcriptional read-in/out. The standard SIN-LTR and no sequence controls were included as positive and negative controls respectively. Luciferase activity was normalised to that of construct 4 (set at 1.0) and data displayed on a log 10 scale (Arbritray units).

FIG. 54. Modelling transcriptional read-in on the bottom strand of an intact ‘supA-2pA-LTR’. As per FIG. 53, supA-LTR sequence was inverted and inserted into a luciferase reporter cassette, except that the inverted RU5 was also included so that an entire supA-/2pA-LTR was present. All constructs 2-9 model the ‘bottom-strand’ orientation (i.e. transcription read-in from downstream cellular promoters). The present construct 1 is identical to construct 1 of FIG. 53 (Example 7); present constructs 2, 4 and correspond to construct 8 in FIG. 53 (Example 7), except that the modified (5′) RU5 or the modified (5′) RU5 containing the optimal GU-rich DSE (‘GU2’) has been included where indicated to assess impact of these sequences; and present constructs 3, 5 and 9 correspond to construct 9 in FIG. 53 (Example 7), except that the modified (5′) RU5 or the modified (5′) RU5 containing the optimal GU-rich DSE (‘GU2’) has been included where indicated to assess impact of these sequences. To provide a GU-rich DSE for the early SV40 polyA sequence, the native SV40 early polyA sequence was simply extended to include the native GU-rich DSE (constructs 6-9). Note the GU box is ‘GU-1’ as noted in Table 2. These sequences were inserted into the luciferase polyA reporter and suspension (serum-free) HEK293T cells transfected, followed by luciferase assay of cell lysates to measure transcriptional read-in/out. Luciferase activity was normalised to that of construct 2 (set at 1.0) and data displayed on a log 10 scale (arbitrary units).

FIG. 55. Detailed schematic of example supA-2pA-LTR as part of LV expression cassette and resultant LTR in target cells. The sequences displayed conform to SEQ ID No: 200 [A] and 201 [B], as shown in Table 9. [A] gives the 5′R-to-PBS, and 3′ppt-to-R-embedded heterologous polyadenylation sequence (in this case the SV40 polyA), with the LV backbone and transgene sequences ‘abbreviated’ in between (no promoter sequence driving the cassette is shown). Grey features are typical HIV-1 based LV sequences of RU5 regions, the PBS and 3′ppt. The attachment sites (for integration) are shown in black (‘att’). Modified sequences of the invention are shown in white features, the direction of which are indicated by arrows, showing whether the sequence functions on the top strand (pointing rightwards) or on the bottom strand (pointing leftwards). PolyA signals (pAS), Upstream enhancer (USE), polyA cleavage zone/region (pAn Clv Zone) and downstream enhancer/GU-rich box (DSE, GU) indicated. [B] displays the resultant LTR in target cells, after the completion of reverse transcription/cDNA synthesis; only one LTR is shown for simplicity, since both LTRs flanking the LV will be identical. The same features convention is used as per panel [A].

FIG. 56. Further modelling transcriptional read-in on the bottom strand of a ‘supA-2pA-LTR’; evaluating alternative GU boxes. As per FIG. 54, supA-LTR sequence was inverted and inserted into a luciferase reporter cassette, so that an entire supA-/2pA-LTR was present. All constructs 2-19 model the ‘bottom-strand’ orientation (i.e. transcription read-in from downstream cellular promoters). Bottom strand variants included seven different GU boxes from difference sources as denoted in Table 2, and the inverted RU5 variably encoded the GU2 modification where indicated. These sequences were inserted into the luciferase polyA reporter and suspension (serum-free) HEK293T cells transfected, followed by luciferase assay of cell lysates to measure transcriptional read-in/out. Luciferase activity was normalised to that of construct 2 (set at 1.0) and data displayed on a log 10 scale (arbitrary units).

FIG. 57. Assessing the impact of the inverted GU boxes on ‘top strand’ transcriptional termination. The supA-2pA-LTR variants described in Table 10 and tested in the inverted orientation (i.e. bottom strand termination) in FIG. 56, were flipped within the luciferase reporter cassettes so that these novel LTRs could be assessed for transcriptional termination efficiency in the forward direction (i.e. on the top strand). This allowed the assessment of any impact of the inverted GU boxes (servicing the SV40 early polyA sequence on the bottom strand) on top strand transcriptional termination. These variants were compared to a standard SIN-LTR (construct 1 and 2) or optimal supA-LTR (construct 5 and 6) with or without native HIV-1 polyA signal, respectively. These sequences were inserted into the luciferase polyA reporter and suspension (serum-free) HEK293T cells transfected, followed by luciferase assay of cell lysates to measure transcriptional read-in/out. Luciferase activity was normalised to that of construct 3 (set at 1.0) and data displayed on a log 10 scale (arbitrary units).

FIG. 58. Production of lentiviral vectors containing variant supA-2pA-LTRs. The configurations of 5′ and 3′ LTRs (for production expression cassettes) are indicated with or without different elements of the invention; all constructs were driven by the CMV promoter (with the ‘3G’ TSS) and the ‘back-up’ polyadenylation sequence was absent except for the standard/MSD controls using a standard SIN-LTR (had SV40 late polyA downstream of the 3′SIN-LTR—not shown). Mutated native HIV-1 5′ PAS (5′pA) and/or major splice donor site (MSD) are indicated by a cross. An internal EFS-GFP-wPRE cassette was present (not shown). The 5′ R region was either the wild type/standard TAR/SL1 or the modified 5′ R SL1 comprised the 3GR-GU2 variant. The heterologous SV40 late polyadenylation signal was positioned downstream of the 3′ppt/ΔU3 region. R region sequences composed of up to 20 nucleotides of HIV-1 R region were inserted between the heterologous PAS and the cleavage site/GU-rich DSE element. The ‘3G-R20’ sequence is the same as the ‘R.1-20’ sequence referred to herein elsewhere. The optional presence of the inverted GU boxes (iGU-1 to iGU-7; see Table 10) are noted, to provide a DSE for the inverted polyA sequences. The structure of the final SIN/supA/supA-2pA LTRs in the ‘target’ cell are provided. LVs were produced in serum-free, suspension HEK293T cells as described elsewhere in the invention, with p256U1 provided in trans for the MSD-mutated LVs. Vector supernatants were titrated on adherent HEK293T cells, followed by flow cytometry (GFP positive cells) and integration assay, and data plotted on a log 10 scale.

FIG. 59. Production of different types of lentiviral vectors containing a supA-2pA-LTR. The configurations of 5′ and 3′ LTRs are indicated with or without different elements of the invention; all constructs were driven by the CMV promoter (with the ‘3G’ TSS) and the ‘back-up’ polyadenylation sequence was absent except for the standard/MSD controls using a standard SIN-LTR (had SV40 late polyA downstream of the 3′SIN-LTR—not shown). Mutated native HIV-1 5′ PAS (5′pA) and/or major splice donor site (MSD) are indicated by a cross. An internal EF1a- or huPGKpromoter driven GFP cassette was present. The 5′ R region was either the wild type/standard TAR/SL1 or the modified 5′ R SL1 comprised the 3GR-GU2 variant. The heterologous SV40 late polyadenylation signal was positioned downstream of the 3′ppt/ΔU3 region. R region sequences composed of up to 20 nucleotides of HIV-1 R region were inserted between the heterologous PAS and the cleavage site/GU-rich DSE element. The ‘3G-R20’ sequence is the same as the ‘R.1-20’ sequence referred to herein elsewhere. The inverted GU box ‘GU-7’ (see Table 10) was used, to provide a DSE for the inverted polyA sequences. LVs were produced in serum-free, suspension HEK293T cells, with p256U1 provided in trans for the MSD-mutated LVs. Vector supernatants were titrated on adherent HEK293T cells, followed by flow cytometry (GFP positive cells) and integration assay, and data plotted on a log 10 scale.

FIG. 60. Measuring read-through the 5′LTR of integrated LVs bearing SIN-, supA or supA-2pA-LTRs. LVs produced in Example 26 (FIG. 58) were used to transduce adherent HEK293T cells or primary donkey fibroblasts (92BR) at MOI 1, followed by passaging for 10 days and integration assay to obtain vector-copy number (HIV Psi qPCR). Total RNA was extracted and HIV-Psi RNA and GAPDH mRNA quantified by RT-qPCR, and a relative HIV-Psi RNA ratio to GAPDH mRNA generated to provide a measure of mobilised HIV-Psi RNA (i.e. read-through the 5′LTR) in each culture. This value was then divided by average copy number per cell, which ranged from 0.8 to 3.6 copies across both cell types.

FIG. 61. Comparison of transcriptional read-in from chromatin upstream of 5′LTR into integrated LV cassettes bearing either standard SIN-LTRs or supA(2pA)-LTRs.

Configuration of integrated LVs bearing either SIN-LTRs or supA(2pA)-LTRs, with optional mutation of the major splice donor (both types) and/or optional mutation of the native HIV-1 pA signal (supA(2pA)-LTR only). LVs were used to transduce adherent HEK293T cells at MOI of 1, and after a 10 day passage host cell genomic DNA extracted for integration assay to determine vector copy number (qPCR to HIV-Psi). PolyA-selected RNA was purified and subjected to RNAseq. Read coverage was mapped to templates for the integrated cassette for each genome. Read counts from regions indicated (MSD[core-Psi], Gag-Psi, and RRE) were initially normalised to total read counts across the GFP transgene. The data were further normalised to vector copy number. Data were finally expressed as % of the MSD reads-depth of the control genome (STD-LV(MSD+)-SIN). Fold-reduction in detected read-through RNA (relative MSD reads of STD-LV(MSD+)-SIN control) is tabulated (LOD=limit of detection).

FIG. 62. Vector-Intron LVs harbouring self-cleaving elements within the transgene 3′UTR: use of production cell derived microRNA target sites. The inverted transgene cassette comprises self-cleaving elements within the 3′UTR sequence that is encompassed by the Vector-Intron sequence on the top strand, and thus such elements are spliced out of packaged vRNA and not delivered to target cells. Self-cleaving elements (such as ribozymes [Z]) eliminate transgene mRNA, and therefore avoid triggering dsRNA-sensing pathways that otherwise reduce LV titres, as well as leading to suppression of transgene protein expression that might otherwise impact on LV titres. In this case, one or more microRNA target sequences are inserted into the 3′UTR, optionally with other self-cleaving elements such as ribozymes. These target sequences may be synthetic, and be targeted by a miRNA expressed exogenously (e.g. by a U6-driven cassette introduced into the production cell) or by endogenous miRNAs.

FIG. 63. Production cell transgene expression and output titres of Vector-Intron LVs harbouring self-cleaving elements within the transgene 3′UTR: use of production cell derived microRNA target sites. Vector-Intron LVs harbouring an inverted EF1a-GFP cassette were generated in a similar format as per FIG. 62. Specifically, the 3′UTR of the inverted transgene that is encompassed by the VI on the top strand had 1× or 3× copies of three different target sequences of miRNAs found to be endogenously expressed in HEK293(T) cells (miR17-5p, miR20a and mi106a). Two sets of variants were produced in which the ribozymes T3H38 and HDV_AG were additionally present within the VI-encompassed 3′UTR region (at positions [1] and [2] respectively). For the variants containing both ribozymes and miRNA target sequence(s), the miRNA target sequences were positioned between the two ribozymes. A third variant type was generated in which a single copy of all three miRNAs were present between the ribozymes (17-5p/20a/106a). LVs were produced in suspension (serum-free) HEK293T cells alongside a standard LV, containing the EF1a-GFP cassette in the forward orientation. Post-production cells were analysed by flow cytometry to generate GFP Expression scores (% GFP×MFI), and resultant vector supernatants were titrated on adherent HEK293T cells by flow cytometry to yield GFP TU/mL values. Titre values and GFP expression scores were normalised to that attained by the standard LV (set to 100%).

FIG. 64. Vector-Intron LVs harbouring self-cleaving elements within the transgene 3′UTR: use of Vector-Intron embedded microRNAs. The figure displays a similar LV production system to that described in FIGS. 44 and 45. The inverted transgene cassette comprises self-cleaving elements within the 3′UTR sequence that is encompassed by the Vector-Intron sequence on the top strand, and thus such elements are spliced out of packaged vRNA and not delivered to target cells. Self-cleaving elements (such as ribozymes [Z]) eliminate transgene mRNA, and therefore avoid triggering dsRNA-sensing pathways that otherwise reduce LV titres, as well as leading to suppression of transgene protein expression that might otherwise impact on LV titres. In this case, one or more microRNA cassettes are inserted into the 3′UTR (processing of which will cleave the transgene mRNA), and optionally the miRNAs produced from processing target sites within the transgene mRNA (in this case 3′UTR sequence). Optionally, these miRs/miRNA targets are combined with other self-cleaving elements such as ribozymes.

FIG. 65. Transgene expression levels in suspension Jurkat cells (T-cell line) transduced with RRE/rev-dependent lentiviral vectors harbouring different 3′ UTR cis-acting elements at matched MOIs (diagonal lines). LV-RRE-EFS-GFP vector stocks produced in suspension (serum-free) HEK293 Ts were initially titrated on adherent HEK293T cells to generate integrating titres (open bars; TU/mL). Vector stocks were used to transduce fresh a Jurkat cells at matched multiplicity of infection (MOI): MOI 1 or 0.5. Different 3′ UTR cis-acting elements were employed as ‘stand-alone’ elements: wPRE, 16×10 bp CARe tiles (CARe.16t) or a single copy of the ZCCHC14 stem loop (HCMV.ZSL1), compared to no element (ΔwPRE). Additionally, variants deleted for wPRE but containing a single copy of the ZCCHC14 stem loop (at position 2) were also paired with increasing numbers of CARe tile, from 1× to 20×10 bp copies (at position 1 i.e. upstream of position 2). Transgene (GFP) expression in transduced Jurkat cells was measured by flow cytometry three days post-transduction and median fluorescence intensities (Arbitrary units) normalised to that achieved with the standard wPRE-containing LV (set to 100%).

FIG. 66. Transgene expression levels in suspension Jurkat cells (T-cell line) transduced with RRE/rev-dependent lentiviral vectors harbouring different 3′ UTR cis-acting elements at matched MOI. LV-RRE-EFS-GFP vector stocks produced in suspension (serum-free) HEK293 Ts were initially titrated on adherent HEK293T cells to generate integrating titres (not shown). Vector stocks were used to transduce fresh Jurkat cells at matched multiplicity of infection of 1. Different 3′ UTR cis-acting elements were employed as ‘stand-alone’ elements: wPRE (black bar), 16×10 bp CARe tiles (CARe.16t; dark grey bar) or a single copy of the ZCCHC14 stem loop (HCMV.ZSL1; striped light grey bar), compared to no element (ΔwPRE; white bar). Additionally, variants deleted forwPRE but containing a single copy of the ZCCHC14 stem loop (at position 2) were also paired with 16×10 bp CARe tiles that contained synthetic variant sequences of the consensus (CARe.16t_vX; at position 1 i.e. upstream of position 2) as shown (grey bars). Transgene (GFP) expression in transduced Jurkat cells was measured by flow cytometry ten days post-transduction and median fluorescence intensities (Arbitrary units) normalised to vector copy-number (VCN), which was measured by qPCR against HIV-Psi on extracted host cell DNA. The solid horizontal line indicates expression level achieved by the larger wPRE element, and the dotted horizontal line indicates expression levels without any 3′UTR element.

FIG. 67. Transgene expression levels in suspension Jurkat cells (T-cell line) transduced with RRE/rev-dependent lentiviral vectors harbouring different 3′ UTR cis-acting elements at matched MOI. LV-RRE-EFS-GFP vector stocks produced in suspension (serum-free) HEK293 Ts were initially titrated on adherent HEK293T cells to generate integrating titres (not shown). Vector stocks were used to transduce fresh Jurkat cells at matched multiplicity of infection of 1. Different 3′ UTR cis-acting elements were employed as ‘stand-alone’ elements: wPRE (black bar), 16×10 bp CARe tiles (CARe.16t; dark grey bar) or a single copy of the ZCCHC14 stem loop (HCMV.ZSL1; striped light grey bar), compared to no element (ΔwPRE; white bar). Additionally, variants deleted forwPRE but containing a single copy of the ZCCHC14 stem loop (at position 2) were also paired with 16×10 bp CARe tiles that contained native variant sequences of the consensus (CARe.16t_vX; at position 1 i.e. upstream of position 2) as shown (grey bars). These were from c-Jun, HSPB3, IFN-alpha and IFN-beta mRNAs. Transgene (GFP) expression in transduced Jurkat cells was measured by flow cytometry ten days post-transduction and median fluorescence intensities (Arbitrary units) normalised to vector copy-number (VCN), which was measured by qPCR against HIV-Psi on extracted host cell DNA. The solid horizontal line indicates expression level achieved by the larger wPRE element, and the dotted horizontal line indicates expression levels without any 3′UTR element.

FIG. 68. Example rAAV vector genomes containing no (empty) or 3′UTR elements to enhance transgene expression in target cells. The ‘CAZL’ element (a composite of tandem CARe 10 bp consensus tiles [CARe.xt] and the ZCCHC14 stem loop [ZSL1]) is ˜140-260 nts in length (depending on use of 4× to 16×CARe tiles). The wPRE is ˜590 nts in length, and therefore occupies more of the rAAV vector genome, size being a critical limitation for rAAVs. Key—Inverted terminal repeat (ITR), Promoter (Pro), Gene of interest (GOI), polyadenylation signal (polyA).

FIG. 69. The results of an experiment wherein rAAV vectors containing either CMV- or EFS-promoter driven GFP, paired with either variant CARe/ZSL1 (‘CAZL’) elements from ˜140-to-260 nts in length or with the wPRE (˜590 nts) at position ‘x’ (i.e. in 3′UTR). Controls for the CARe/ZSL1 were inverted elements (‘inv’) to control for potential effects of different genome sizes, which ranged from 2.0-to-2.6 kb (CMV) and 1.6-to-2.2 kb (EFS) across all genomes. An empty rAAV vector was used as a negative control. rAAVs were made by co-transfection of HEK293T suspension cells with pGenome/pRepCap/pHelper plasmids at 1:1:1 ratio, and harvest 72 hours post-transfection. Vector harvest material was titrated by qPCR against the GFP sequence to generated vg/mL physical titre values. HEPG2 cells were transduced at the denoted MOIs, and then 72 hours post-transduction cells analysed by flow cytometry, and GFP Expression scores (% GFP+ x MFI; ArbUs) generated.

FIG. 70. High titre production of an LV encoding a chimeric antigen receptor (CAR) transgene cassette using an optimised Vector-Intron efficiently spliced out of packaged vRNA. [A] RT-PCR analysis of vRNA-derived species within production cells (total and cytoplasmic) and resultant LV virions (V) for four different types of LV vector backbone expression cassettes. ‘STD’ refers to standard 3^rdGen LVs, harbouring all the typical cis-acting elements, including the major splice donor (MSD) and rev-response element (RRE). ‘2KO’ refers to newer generation LVs wherein the MSD has been mutated; these also contain the RRE and require expression of a modified U1 snRNA (256U1) molecule to fully restore output titres. 256U1 used with STD LVs also increases packaged vRNA and titres, see [B]. The ‘MaxPax’ LV contains the v5.6 variant of the Vector-Intron of the present invention, and harbours both mutation in the MSD and deletion of the RRE and gag-p17INS sequences. Promoters used were either EF1a or the short EF1a (EFS). The RT-PCR use primers upstream of the MSD (fwd) and downstream within the GFP transgene so that aberrant or ‘correct’ VI splicing could be monitored. The presence of packageable/packaged vRNA is shown (ψ-vRNA); note that the size of RT-PCR product reflects the size of promoter (EFS is ˜1kb shorter than EF1a; see 2KO-EF1a vs 2KO-EFS) and the increase in capacity of the VI-derived vRNA of ˜1kb (see 2KO-EFS vs MaxPax-EFS). An RT-PCR to actin mRNA was used as a positive control. Panel B displays the output titres of the vectors descrive in A; both integration and ‘biological’ (scFV) titres are shown against the assay reference control (stripes). VI-derived MaxPax LVs were produced in the absence of rev.

FIG. 71. Production titres of standard LV-GFPs harbouring different transgene promoters, SIN or supA-2pA LTRs, and with alternative 3′UTR elements. The figure displays the structure of the LV DNA expression cassette in each case, from 5′ to 3′ LTR. At the 5′ LTR position the CMV promoter is used with the 3Gs at the transcription start site (CMV-3G), has either the standard R region (TAR) or the supA(2pA)-ITR 5′ modification ‘GU2’, and is optionally mutated for the native 5′ polyA (white X). All standard LVs contained an intact major splice donor (MSD) and rev response element (RRE). The GFP transgene was driven by either EF1a, EFS (short EF1a i.e. lacking intron A) or human PGK promoters. The 3′ UTR was the either the wPRE or a CARe/ZSL1 element containing 8× tiles of the CARe 10 bp consensus element link to a ZCCHC14 protein-binding stem loop, or did not contain an element (white X). The 3′ LTR were either standard, self-inactivating (SIN) or used the 3′ R-embedded heterologous polyA adenylation sequence (in this case SV40 late polyA), with inverted polyA (not shown; in this case the SV40 early polyA) and GU-box (iGU7; in this case from the ‘SPA’ polyA, based on beta-globin polyA) upstream of the Usptream enhancer (USE; in this case from SV40 late polyA). Note that constructs using the SIN-LTR also utilised a ‘back-up’ polyA downstream (in this case the SV40 pA), wherease supA(2p)-LTRs did not. The embedded R region was downstream of the heterologous polyA signal (PAS) and comprised the first 20 nucleotides of the R region, including the 3Gs (3G-R20). The heterologous polyA cleavage zone and downstream DSE/GU rich sequence are also indicated (clv-DSE/GU). LVs were produced in suspension (serum-free) HEK293T cells and titrated on adherent HEK293T cells by flow cytometry (GFP-FACS; grey bars) and by integration assay (qPCR for HIV-Psi, on host cell DNA; black bars) on days 3 and 10 post-transduction.

FIG. 72. Production titres of ‘2KO’-LV-GFPs harbouring different transgene promoters, SIN or supA-2pA LTRs, and with alternative 3′UTR elements. The figure displays the structure of the LV DNA expression cassette in each case, from 5′ to 3′ LTR. At the 5′ LTR position the CMV promoter is used with the 3Gs at the transcription start site (CMV-3G), has either the standard R region (TAR) or the supA(2pA)-ITR 5′ modification ‘GU2’, and is optionally mutated for the native 5′ polyA (white X). All 2KO-LVs contained a mutated major splice donor (MSD; mutant ‘2KOm5’ was used) and rev response element (RRE). The GFP transgene was driven by either EF1a, EFS (short EF1a i.e. lacking intron A) or human PGK promoters. The 3′ UTR was the either the wPRE or a CARe/ZSL1 element containing 8× tiles of the CARe 10 bp consensus element link to a ZCCHC14 protein-binding stem loop, or did not contain an element (white X). The 3′ LTR were either standard, self-inactivating (SIN) or used the 3′ R-embedded heterologous polyA adenylation sequence (in this case SV40 late polyA), with inverted polyA (not shown; in this case the SV40 early polyA) and GU-box (iGU7; in this case from the ‘SPA’ polyA, based on beta-globin polyA) upstream of the Usptream enhancer (USE; in this case from SV40 late polyA). Note that constructs using the SIN-LTR also utilised a ‘back-up’ polyA downstream (in this case the SV40 pA), wherease supA(2p)-LTRs did not. The embedded R region was downstream of the heterologous polyA signal (PAS) and comprised the first 20 nucleotides of the R region, including the 3Gs (3G-R20). The heterologous polyA cleavage zone and downstream DSE/GU rich sequence are also indicated (clv-DSE/GU). LVs were produced in suspension (serum-free) HEK293T cells and titrated on adherent HEK293T cells by flow cytometry (GFP-FACS; grey bars) and by integration assay (qPCR for HIV-Psi, on host cell DNA; black bars) on days 3 and 10 post-transduction.

FIG. 73. Production titres of ‘MaxPax’ (Vector-Intron) LV-GFPs harbouring different transgene promoters, SIN or supA-2pA LTRs, and with alternative 3′UTR elements. The figure displays the structure of the LV DNA expression cassette in each case, from 5′ to 3′ LTR. At the 5′ LTR position the CMV promoter is used with the 3Gs at the transcription start site (CMV-3G), has either the standard R region (TAR) or the supA(2pA)-ITR 5′ modification ‘GU2’, and is optionally mutated for the native 5′ polyA (white X). All MaxPax-LVs contained a mutated major splice donor (MSD; mutant ‘2KOm5’ was used) and Vector-Intron (VI) variant v5.5, as well as truncated gag-Psi region (not shown). The GFP transgene was driven by either EFS (short EF1a i.e. lacking intron A) or human PGK promoters. The 3′ UTR was the either the wPRE or a CARe/ZSL1 element containing 8× tiles of the CARe 10 bp consensus element link to a ZCCHC14 protein-binding stem loop, or did not contain an element (white X). The 3′ LTR were either standard, self-inactivating (SIN) or used the 3′ R-embedded heterologous polyA adenylation sequence (in this case SV40 late polyA), with inverted polyA (not shown; in this case the SV40 early polyA) and GU-box (iGU7; in this case from the ‘SPA’ polyA, based on beta-globin polyA) upstream of the Usptream enhancer (USE; in this case from SV40 late polyA). Note that constructs using the SIN-LTR also utilised a ‘back-up’ polyA downstream (in this case the SV40 pA), wherease supA(2p)-LTRs did not. The embedded R region was downstream of the heterologous polyA signal (PAS) and comprised the first 20 nucleotides of the R region, including the 3Gs (3G-R20). The heterologous polyA cleavage zone and downstream DSE/GU rich sequence are also indicated (clv-DSE/GU). LVs were produced in suspension (serum-free) HEK293T cells and titrated on adherent HEK293T cells by flow cytometry (GFP-FACS; grey bars) and by integration assay (qPCR for HIV-Psi, on host cell DNA; black bars) on days 3 and 10 post-transduction.

FIG. 74. Transcriptional read-in to the 5′ LTR of integrated standard LV-GFPs harbouring different transgene promoters, SIN or supA-2pA LTRs, and with alternative 3′UTR elements. The figure displays the structure of the integrated LV DNA cassette in each case, from 5′ to 3′ LTR, resulting from reverse transcription and integrated of the LVs in FIG. 71. Therefore, both 5′ and 3′ LTRs are identical. The standard SIN LTR contains the deleted U3 region (ΔU3), and R-U5, which contains the native HIV-1 polyA signal. The supA(2pA) LTRs harbour an inverted polyA (not shown; in this case the SV40 early polyA) and GU-box (iGU7; in this case from the ‘SPA’ polyA, based on beta-globin polyA) upstream of the Usptream enhancer (USE; in this case from SV40 late polyA). Downstream of the USE is the heterologous polyA signal (SV40 late PAS), the GU2-modified R region, and is optionally mutated for the native HIV-1 polyA signal in the U5 region (white X). All other aspects between the LTRs were the same as described in FIG. 71. LVs were used to transduce adherent HEK293T cells at MOI 1, followed by passaging for 10 days and integration assay to obtain vector-copy number (HIV Psi qPCR). Total RNA was extracted, and HIV-Psi RNA and GAPDH mRNA quantified by RT-qPCR. The relative HIV-Psi RNA ratio to GAPDH mRNA generated to provide a measure of mobilised HIV-Psi RNA (i.e. read-through the 5′LTR) in each culture (normalised to vector copy-number); these are plotted in comparison to read-in values for the LV harbouring the standard wPRE/SIN-LTR variant for each LV bearing the same internal promoter driving the transgene (set to 100%, black bars).

FIG. 75. Transcriptional read-in to the 5′ LTR of integrated ‘2KO’ LV-GFPs harbouring different transgene promoters, SIN or supA-2pA LTRs, and with alternative 3′UTR elements. The figure displays the structure of the integrated LV DNA cassette in each case, from 5′ to 3′ LTR, resulting from reverse transcription and integrated of the LVs in FIG. 72. Therefore, both 5′ and 3′ LTRs are identical. The standard SIN LTR contains the deleted U3 region (ΔU3), and R-U5, which contains the native HIV-1 polyA signal. The supA(2pA) LTRs harbour an inverted polyA (not shown; in this case the SV40 early polyA) and GU-box (iGU7; in this case from the ‘SPA’ polyA, based on beta-globin polyA) upstream of the Usptream enhancer (USE; in this case from SV40 late polyA). Downstream of the USE is the heterologous polyA signal (SV40 late PAS), the GU2-modified R region, and is optionally mutated for the native HIV-1 polyA signal in the U5 region (white X). All other aspects between the LTRs were the same as described in FIG. 72, including the mutated MSD. LVs were used to transduce adherent HEK293T cells at MOI 1, followed by passaging for 10 days and integration assay to obtain vector-copy number (HIV Psi qPCR). Total RNA was extracted, and HIV-Psi RNA and GAPDH mRNA quantified by RT-qPCR. The relative HIV-Psi RNA ratio to GAPDH mRNA generated to provide a measure of mobilised HIV-Psi RNA (i.e. read-through the 5′LTR) in each culture (normalised to vector copy-number); these are plotted in comparison to read-in values for the LV harbouring the standard wPRE/SIN-LTR variant for each LV bearing the same internal promoter driving the transgene (set to 100%, black bars).

FIG. 76. Transcriptional read-in to the 5′ LTR of integrated ‘MaxPax’ (Vector-Intron) LV-GFPs harbouring different transgene promoters, SIN or supA-2pA LTRs, and with alternative 3′UTR elements. The figure displays the structure of the integrated LV DNA cassette in each case, from 5′ to 3′ LTR, resulting from reverse transcription and integrated of the LVs in FIG. 73. Therefore, both 5′ and 3′ LTRs are identical. The standard SIN LTR contains the deleted U3 region (ΔU3), and R-U5, which contains the native HIV-1 polyA signal. The supA(2pA) LTRs harbour an inverted polyA (not shown; in this case the SV40 early polyA) and GU-box (iGU7; in this case from the ‘SPA’ polyA, based on beta-globin polyA) upstream of the Usptream enhancer (USE; in this case from SV40 late polyA). Downstream of the USE is the heterologous polyA signal (SV40 late PAS), the GU2-modified R region, and is optionally mutated for the native HIV-1 polyA signal in the U5 region (white X). All other aspects between the LTRs were the same as described in FIG. 73, including the mutated MSD and replacement of RRE with VI, and truncated gag-Psi (not shown). LVs were used to transduce adherent HEK293T cells at MOI 1, followed by passaging for 10 days and integration assay to obtain vector-copy number (HIV Psi qPCR). Total RNA was extracted, and HIV-Psi RNA and GAPDH mRNA quantified by RT-qPCR. The relative HIV-Psi RNA ratio to GAPDH mRNA generated to provide a measure of mobilised HIV-Psi RNA (i.e. read-through the 5′LTR) in each culture (normalised to vector copy-number); these are plotted in comparison to read-in values for the LV harbouring the standard wPRE/SIN-LTR variant for each LV bearing the same internal promoter driving the transgene (set to 100%, black bars).

FIG. 77. Relative transgene expression of integrated standard LV-GFPs harbouring different transgene promoters, SIN or supA-2pA LTRs, and with alternative 3′UTR elements. The figure re-states the configuration of the integrated LV as described in FIG. 74. Adherent HEK293T cells transduced with the LVs at MOI of 1 (as per FIGS. 71 and 74) were analysed by flow cytometry at day 3 post-transduction and GFP Expression scores generated by multiplying % GFP-positive cells and median fluorescence values (MFI). These ES scores were divided by the vector-copy number generated at day 10 post-transduction (by qPCR to HIV Psi). The resulting ‘relative GOI Exprn’ values are plotted in comparison to those of the LV harbouring the standard wPRE/SIN-LTR for each LV bearing the same internal promoter driving the transgene (set to 100, black bars).

FIG. 78. Relative transgene expression of integrated standard LV-GFPs harbouring different transgene promoters, SIN or supA-2pA LTRs, and with alternative 3′UTR elements. The figure re-states the configuration of the integrated LV as described in FIG. 75. Adherent HEK293T cells transduced with the LVs at MOI of 1 (as per FIGS. 72 and 75) were analysed by flow cytometry at day 3 post-transduction and GFP Expression scores generated by multiplying % GFP-positive cells and median fluorescence values (MFI). These ES scores were divided by the vector-copy number generated at day 10 post-transduction (by qPCR to HIV Psi). The resulting ‘relative GOI Exprn’ values are plotted in comparison to those of the LV harbouring the standard wPRE/SIN-LTR for each LV bearing the same internal promoter driving the transgene (set to 100, black bars).

FIG. 79. Relative transgene expression of integrated standard LV-GFPs harbouring different transgene promoters, SIN or supA-2pA LTRs, and with alternative 3′UTR elements. The figure re-states the configuration of the integrated LV as described in FIG. 76. Adherent HEK293T cells transduced with the LVs at MOI of 1 (as per FIGS. 73 and 76) were analysed by flow cytometry at day 3 post-transduction and GFP Expression scores generated by multiplying % GFP-positive cells and median fluorescence values (MFI). These ES scores were divided by the vector-copy number generated at day 10 post-transduction (by qPCR to HIV Psi). The resulting ‘relative GOI Exprn’ values are plotted in comparison to those of the LV harbouring the standard wPRE/SIN-LTR for each LV bearing the same internal promoter driving the transgene (set to 100, black bars).

FIG. 80. Use of the CARe/ZSL1 3′UTR element within the inverted transgene cassette of a Vector-Intron genome with ‘functionalised 3′UTR’. The figure provides an alternative structure to that of FIGS. 68 and 69. The inverted transgene configuration within a Vector-Intron LV is desirable to enable on-boarding of intron-containing cassettes. The functionalised 3′UTR in this instance contains two self-cleaving ribozymes (Z). In this case, the transgene is used with the CARe.8t/ZSL1 element (CAZL), which is positioned outside of the VI-excised sequence, and so will remain in the delivered LV transgene cassette.

FIG. 81. Use of the CARe/ZSL1 3′UTR element within the inverted transgene cassette of a Vector-Intron genome with ‘functionalised 3′UTR’ improves output titres. The Vector-Intron LVs based on FIG. 80 were made +/−the CARe/ZSL1 3′UTR element (aka ‘CAZL’ and +/−the dual ribozymes functionalising the 3′UTR within the VI encoded on the top strand. LVs were produced in suspension (serum-free) HEK293T cells and post-production levels of GFP expression measured by flow cytometry (% GFP×MFI). Vector supernatants were titrated on adherent HEK293T cells by flow cytometry and integration assay (qPCR to HIV Psi) at day 3 and 10 post-transduction respectively. Data was normalised to a standard LV containing a forward facing transgene cassette (EF1a-GFP) and the wPRE (set to 100%).

DETAILED DESCRIPTION OF THE INVENTION
Nucleotide Sequence and Set of Nucleotide Sequences

The present inventors surprisingly found that:

- 1) employing modified polyadenylation (polyA) sequences within LV genome expression cassettes results in simplified production of vector genomic RNA for packaging, improved transgene expression and reduced transcriptional read-in and -out (both of the vector genome expression cassette and transgene expression cassette) in transduced cells;
- 2) viral vectors with novel short cis-acting sequences in the 3′ UTR of a transgene expression cassette either in addition to traditional PREs to boost transgene expression in target cells or to replace these longer PREs entirely, enabling increased transgene capacity whilst maintaining high levels of transgene expression in target cells;
- 3) introduction of an intron into the vector genome expression cassette facilitates removal of the rev-response element (RRE), which allows for more transgene capacity in the vector; and
- 4) RNAi can be employed in retroviral vector production cells to suppress the expression of the NOI (i.e. transgene) during retroviral vector production in order to minimize unwanted effects of the transgene protein and to rescue of titres of retroviral vectors harbouring an actively transcribed inverted transgene cassette (wherein the transgene expression cassette is all or in part inverted with respect to the retroviral vector genome expression cassette).

In some embodiments, the lentiviral vector genome expression cassette comprises a transgene expression cassette.

In some embodiments, the lentiviral vector genome expression cassette comprises a transgene expression cassette and a vector intron.

In some embodiments, the vector intron comprises one or more transgene mRNA self-destabilization or self-decay element(s) or one or more transgene mRNA nuclear retention signal(s).

In a further aspect, the invention provides a nucleotide sequence comprising a transgene expression cassette wherein the 3′ UTR of the transgene expression cassette comprises at least one cis-acting sequence selected from (a) a cis-acting Cytoplasmic Accumulation Region (CAR) sequence; and/or (b) a cis-acting ZCCHC14 protein-binding sequence.

In some embodiments, the major splice donor site in the lentiviral vector genome expression cassette is inactivated.

In some embodiments, the lentiviral vector genome expression cassette does not comprise a rev-response element (RRE).

In some embodiments, the cryptic splice donor site adjacent to the 3′ end of the major splice donor site in the lentiviral vector genome expression cassette is inactivated.

In some embodiments, the transgene expression cassette is in the forward orientation with respect to the lentiviral vector genome expression cassette. Thus, the transgene expression cassette and vector intron may not be transcriptionally opposed.

In some embodiments, the transgene expression cassette is inverted with respect to the lentiviral vector genome expression cassette. Thus, the transgene expression cassette and vector intron are transcriptionally opposed.

In some embodiments:

- a) the vector intron is not located between the promoter of the transgene expression cassette and the transgene; and/or
- b) the nucleotide sequence comprises a sequence as set forth in any of SEQ ID NOs: 2, 3, 4, 6, 7, and/or 8, and/or the sequences CAGACA, and/or GTGGAGACT; and/or
- c) the 3′ UTR of the transgene expression cassette comprises the vector intron.

In some embodiments, the vector intron comprises one or more transgene mRNA self-destabilization or self-decay element(s) or one or more transgene mRNA nuclear retention signal(s).

In some embodiments, the nucleotide sequence comprises a lentiviral vector genome expression cassette, wherein:

- i) the major splice donor site and cryptic splice donor site adjacent to the 3′ end of the major splice donor site in the lentiviral vector genome expression cassette are inactivated;
- ii) the lentiviral vector genome expression cassette does not comprise a rev-response element;
- iii) the lentiviral vector genome expression cassette comprises a transgene expression cassette and a vector intron; and
- iv)
  - a) When the transgene expression cassette is inverted with respect to the lentiviral vector genome expression cassette:
    - i. the vector intron is not located between the promoter of the transgene expression cassette and the transgene; and
    - ii. the nucleotide sequence comprises a sequence as set forth in any of SEQ ID NOs: 2, 3, 4, 6, 7, and/or 8, and/or the sequence CAGACA, and/or GTGGAGACT; and
    - iii. the 3′ UTR of the transgene expression cassette comprises the vector intron; and
  - b) the vector intron comprises one or more transgene mRNA self-destabilization or self-decay element(s) or one or more transgene mRNA nuclear retention signal(s).

As mentioned above, any one or more of the aspects of the invention described herein may be combined. This provides the advantage that the surprising and beneficial effects of each aspect can be achieved in combination, i.e. the inclusion of each aspect has an additive and/or synergistic effect.

Suitably, any two of the aspects of the invention described herein may be combined. Suitably any three of the aspects of the invention described herein may be combined. Suitably, any four of the aspects of the invention may be combined. Suitably, all aspects of the invention described herein may be combined. Therefore, all of the embodiments of the invention described herein with respect to one aspect of the invention also relate to any and all other aspect(s) of the invention.