FIELD OF THE INVENTION

The present invention relates to a process for the production of recombinant lentiviral vectors. The described method combines the use of a rocking motion bioreactor, pH adjustment and harvest by perfusion to provide advantageously high yields. Accompanying methods of lentiviral vector particle purification are also described.

BACKGROUND TO THE INVENTION

Recombinant viral vectors have become an important tool for cell therapy applications, such as adoptive cell therapies using genetically engineered T cells carrying a modified T cell receptor or chimeric antigen receptor (CAR). In particular, lentiviral vectors (part of the retrovirus family) are useful due to their ability to transduce and integrate into the genome of both dividing and non-dividing cells.

Vector particles may be produced by the transient transfection of human cells, such as human embryonic kidney (HEK) 293T cells, with plasmids encoding the viral genome components and the transgene of interest. The cells must then be cultured under suitable conditions for optimal production of the viral particles.

The resulting viral particles must then also be purified before being used to transduce immune cells. This purification process typically leads to loss of some viral particles and must be optimized to give acceptable yields.

There is therefore a need in the art for improved methods of production and purification of lentiviral vectors. SUMMARY OF THE INVENTION

The present inventors have developed a method for the production of recombinant lentiviral vectors using a combination of cell culture in a rocking-motion bioreactor and harvesting by perfusion. This method results in improved yield compared to prior art methods. Furthermore, the method may also include downstream purification steps adapted to the production method which further improve the final yield of purified lentiviral vectors.

In a first aspect, the present invention provides a method for the production of a recombinant lentiviral vector, comprising:

(a) inoculating serum-free media in a rocking motion bioreactor with human cells;

(b) expanding the cells for from 1 to 4 days;

(c) transfecting the cells with at least one plasmid adapted for the production of a lentiviral vector;

(d) inducing the cells with an induction agent;

(e) adjusting the pH of the media to from 6.2 to 7.0, preferably to from 6.4 to 6.8, most preferably to 6.8;

(f) expanding the cells for from 2 to 8 days while harvesting the produced recombinant lentiviral vector from the culture medium by perfusion through a perfusion membrane; and

(g) optionally purifying the recombinant lentiviral vector.

The pore size of the perfusion membrane used for harvesting may be selected to optimise lentiviral vector recovery. Preferred pore sizes include 1.0, 1.1, 1.2, 1.3, 1.4, 1.5, 1.6, 1.7, 1.8, 1.9, 2.0, 2.1, 2.2, 2.3, 2.4, and 2.5 pm. A particularly preferred pore size is 1.8 pm.

Preferably, the purification in step (g) includes at least one of

(i) endonuclease treatment;

(ii) filtration; or

(iii) tangential flow filtration (TFF).

The filtration step may employ a filter having a pore size of from 0.1 to 0.8 pm. More preferably, the purification step (g) comprises:

(i) endonuclease treatment;

(ii) filtration through a 0.45 pm filter;

(iii) concentration of the media by tangential flow filtration;

(iv) optional buffer exchange; and

(v) filtration through a 0.6/0.2 pm filter.

Preferably the endonuclease treatment is carried out at 4°C for from 6 to 12 hours.

During the transfection step the cells may be transfected with a mixture of polyethyleneimine (PEI) and plasmids.

The induction agent may be sodium butyrate. Preferably the final concentration of sodium butyrate is from 5 to 9 mM. More preferably the final concentration of sodium butyrate is 9 mM.

The pH adjustment in step (e) may be carried out with carbon dioxide or hydrochloric acid, or a combination of carbon dioxide and hydrochloric acid.

The rocking motion bioreactor may be set to either 25 rpm at a 9° rocking angle or 15 rpm at a 12° rocking angle.

The recombinant lentiviral vector may be harvested by perfusion at a perfusion rate of from lOL/day to 40L/day.

The perfusion membrane may have a pore size of from 0.5 pm to 2.5 pm, preferably from 1.0 pm to 2.0 pm, more preferably 1.5 pm to 2.0 pm, most preferably 1.8 pm.

The perfusion membrane may be located in a fixed position relative to the rocking motion bioreactor.

Preferably, steps (a) to (f) are carried out at 37°C. Preferably the cells are HEK293 cells or HEK293T cells. Where HEK293T cells are used, they may be adapted for growth in serum-free media.

DESCRIPTION OF THE FIGURES

Figure 1. Total Cell count and cell culture viability during the expansion phase in Flexsafe RM perfusion bag (Sartorius) and CellBag (GE) prior to transfection.

Cumulative Total cell count (Total cells/ml) in runs 3, 4, 5, 6 and 7 in Flexsafe RM perfusion bag and GE003, GE004 in CellBag (A) and Cumulative Cell viability (%) in runs 3, 4, 5, 6, and 7 in Flexsafe RM perfusion bag and GE003,GE004 in CellBag (B).

Figure 2. Cell growth during expansion and production phases in Flexsafe RM perfusion bag (Sartorius)

Total cell concentration (Total Cells/mL) in runs 002,003,004,005,005,006 and 007 (A), total cell concentration in runs with no pH shift 004,005,007 and in run with pH shift 006 (B). Total cell concertation in runs with no perfusion post-transfection and post-induction; Sar003,004 and with perfusion post-transfection and post induction; 005,006,007 (C).

Figure 3. Cell viability during expansion and production phases in Flexsafe RM perfusion bag (Sartorius) Cell Culture Viability (%) in runs; 002,003,004,005,005,006 and 007 (A), Cell Culture Viability in runs with no pH shift; 004,005,007 and in run with pH shift 006 (B) Cell Culture Viability in runs with no perfusion post-transfection and post-induction; Sar003,004 and with perfusion post-transfection and post-induction; 005,006,007 (C)

Figure 4. Functional titre Harvest I (48 PT) and Harvest II (&2 hrs PT) in Sartorius Flex perfusion bag Comparison of functional particles concertation (TU/ml) produced in bioreactor in runs; 002, 003, 004, 005, 006, and 007(A). Comparison of functional particles between the production bag and collection bags in in runs; 002, 003, 004, 005, 006, and 007 (B).

Figure 5. Physical titre and P:I ratio

Physical titre in runs; 003,004,005,006,007 in production bag and LV collection bag, perfusion effect on physical titre (A), Cumulative data from runs; 003,004,005,006,007 and shake flask control, P:I ratio over time (B)

Figure 6. FACS plots transfection efficiency cell culture transfected in bioreactor and non transfected cells.

The percentage of CAT 19- positive cells determined by flow cytometry represent transfection efficiency for samples harvested 24 hrs post induction in runs;

Sar006 (A) and Sar007 (B) Shake flask CTRL 1 C) and Shake Flask CTRL 2.

Figure 7 MADLS results reporting presence of specific particle populations at each stage of the DSP process according to size (d.nm) and physical particle concentration (pp/mL): a) and b) “Small aggregates”, c) and d) “Vector”, e) and f) “Larger aggregates”. H 24h - Harvest collected after 24h and treated with Denarase®, H 48h - Harvest collected after 48h and treated with Denarase®, TFF-I - clarified Denarase® treated harvest/TFF input, TFF-C R - TFF concentration retentate, TFF-C P - TFF concentration permeate, TFF-D R - TFF diafiltration retentate, TFF-D P - TFF diafiltration permeate, SF - sterile filtered final lentiviral vector product.

Figure 8 Lentivrial recovery obtained by MADLS, p24 ELISA and infectivity assay. TFF- C R - TFF concentration retentate, TFF-C P - TFF concentration permeate, TFF-D R - TFF diafiltration retentate, TFF-D P - TFF diafiltration permeate, SF - sterile filtered final lentiviral vector product.

Figure 9 a) p24 concentration results obtained using ABL inc HIV-1 p24 Antigen Capture Assay, and b) recovery results after mass balancing. H 24h - Harvest collected after 24h and treated with Denarase®, H 48h - Harvest collected after 48h and treated with Denarase®, TFF-I - clarified Denarase® treated harvest/TFF input, TFF-C R - TFF concentration retentate, TFF-C P - TFF concentration permeate, TFF-D R - TFF diafiltration retentate, TFF-D P - TFF diafiltration permeate, SF - sterile filtered final lentiviral vector product.

Figure 10 a) Lentiviral titre (TU/mL) and b) recovery after mass balancing. Infectivity assay was performed only once and 50 000 cells per well were correctly seeded. H 24 - Harvest collected after 24h and treated with Denarase®, H 48 - Harvest collected after 48h and treated with Denarase®, TFF-I - clarified Denarase® treated harvest/TFF input, TFF- C R - TFF concentration retentate, TFF-C P - TFF concentration permeate, TFF-D R - TFF diafiltration retentate, TFF-D P - TFF diafiltration permeate, SF - sterile filtered final lentiviral vector product.

DETAILED DESCRIPTION OF THE INVENTION

1. Vector Production Method

In a first aspect, the present invention provides a method of producing recombinant lentiviral vectors. The present inventors have found that a combination of the use of a rocking motion bioreactor, pH adjustment and harvest by perfusion provides advantageously high yields

The lentiviral particles are produced by cells which have been transfected with one or more plasmids carrying the required genes for the production of virus particles that include the sequence of the desired transgene. In the case of a typical retroviral vector, the viral genes used may include gag (group specific antigen), pol (polymerase), and env (envelope). The gag sequence encodes the three main structural proteins: the matrix protein, nucleocapsid proteins, and capsid protein. The pol sequence encodes the enzymes reverse transcriptase and integrase, the former catalyzing the reverse transcription of the viral genome from RNA to DNA during the infection process and the latter responsible for integrating the proviral DNA into the host cell genome. The env sequence encodes for both SU and TM subunits of the envelope glycoprotein. For lentiviral vectors the rev gene may also be included for regulatory purposes. Additionally, the cells are transfected with a plasmid carrying the transgene of interest and a sequence named packaging signal (y) required for specific packaging of the viral RNA into newly forming virions. This plasmid may also include two LTRs (long terminal repeats), which contain elements required to drive gene expression, reverse transcription and integration into the host cell chromosome. In this way, the transgene of interest is incorporated into the viral particles in a form ready for expression in target cells after transduction. Depending on the exact system used, these genes may be divided onto separate plasmids in order to avoid the inadvertent production of RCLs (replication-competent lentiviruses) via recombination events. Typically, the cells are transiently transfected with plasmids carrying all the required elements so that the cells transiently produce viral particles. Alternatively, certain viral genes may be permanently incorporated into a cell to produce a packaging cell line. Packaging cell lines need only be transfected with the plasmid carrying the transgene of interest and associated packaging signal in order to produce viral particles. Once a cell carries all the elements required for production of viral particles it may be termed a producer cell.

The methods of the present invention are carried out in a rocking motion bioreactor. The term “rocking motion bioreactor” will be understood by those of skill in that art to relate to any device that achieves mixing of a fluid via a rocking or wave motion. Such devices are advantageous because they ensure gas exchange without the high levels of turbulence observed when other mixing methods, such as stirring or shaking, are employed. This is particularly important where the bioreactor is used to culture physically sensitive cells. The wave motion in the fluid is usually achieved by placing the fluid container on a rocking plate. Accordingly, a rocking motion bioreactor may also be referred to as a rocking plate bioreactor, rocking bioreactor or rocker bioreactor, wave bioreactor, or wave-mixed bioreactor. The rate and angle of rocking can be adjusted to control the rate of gas exchange.

The present method typically makes use of serum-free media. Serum-free media is advantageous because the resulting vectors are suitable for therapeutic use and more easily comply with Good Manufacturing Practice (GMP) requirements. Once the media has been inoculated with the desired cells, they must be expanded in order to reach a cell density that will provide the desired quantity of viral particles on an acceptable time scale. In most cases, this will take from 1 to 4 days, depending on the exact conditions and cell type. In the present method, the preferred cells are human embryonic kidney (HEK) 293 cells. HEK293T cells, which have been transformed with the SV40 T-antigen may also be used once adapted for growth in serum-free media. Methods for adapting cells for growth in serum-free media are available to the skilled person. Transfection of the expanded cells is achieved by introduction of plasmids carrying the elements required for lentiviral particle production. Typically transfection is achieved using a mixture of polyethyleneimine (PEI) and the required plasmids. After transfection, viral gene expression is induced via treatment with an induction agent. The induction agent may be sodium butyrate. The final concentration of sodium butyrate may be from 5 to 13 mM. A preferred final concentration of sodium butyrate is 9 mM.

The present method also makes use of a pH adjustment during cell culture. More specifically, the pH of the media is adjusted such that the pH in the range of from 6.2 to 7.0. In particular, the pH of the media may be adjusted such that it is in the range of from 6.4 to 6.8. The preferred pH of the media after adjustment is 6.8. The adjustment of pH may be carried out with carbon dioxide or hydrochloric acid, or a combination of carbon dioxide and hydrochloric acid

After the pH has been adjusted to the desired level the cells are expanded for a period while lentiviral vector particles are produced. This may be from 2 to 8 days and is typically carried out at 37 °C while the rocking motion bioreactor operates to mix the media. Exemplary settings for the rocking motion bioreactor are 25 rpm at a 9° rocking angle or 15 rpm at a 12° rocking angle.

During this time, recombinant lentiviral vector particles are harvested from the culture medium by perfusion through a perfusion membrane. The rate of perfusion must be selected in order to maximise the level of perfusate while maintaining the viability of the cells. The present inventors have found a perfusion rate of from lOL/day to 40L/day to be optimal. Harvesting at 40L/day has been found to be particularly advantageous.

Furthermore, the pore size of the perfusion membrane must be selected to provide optimal results. The present inventors have found pore sizes in the range of from 0.5 pm to 2.5 pm to be useful in the present invention. More specifically, pore sizes in the range of from 1.0 pm to 2.0 pm or from 1.5 pm to 2.0 pm are useful. Preferred pore sizes include 1.0, 1.1, 1.2, 1.3, 1.4, 1.5, 1.6, 1.7, 1.8, 1.9, 2.0, 2.1, 2.2, 2.3, 2.4, and 2.5 pm. The most preferred pore size is 1.8 pm. The perfusion membrane may be located in a fixed position relative to the rocking motion bioreactor. For example, the membrane may be fixed to the bottom of the fluid container. Without wishing to be bound by any one theory, the inventors believe fixing the position of the perfusion membrane in this way reduces turbulence in the reactor and increases cell viability.

2. Vector Purification Method

The present invention also provides a method of vector purification, for use in conjunction with the method of vector production.

The purification will include at least one of the following steps:

(i) endonuclease treatment;

(ii) filtration ; or

(iii) tangential flow filtration (TFF).

Treatment with an endonuclease, for example Benzonase® or Denarase®, removes host cell DNA by digestion. Typically, endonuclease treatment is carried out at from 2 to 8 °C, preferably at 4 °C. The treatment may be carried out overnight or anywhere in the range of from 6 to 12 hours. The amount of endonuclease used can be adjusted depending on the exact enzyme used. For Denarase a typical amount would be 2.5 U/mL.

Filtration of the perfusate removes cell debris and other material. The present inventors have found the use of a 0.45 pm filter to be preferable. In some cases, an additional filtration step at the end of purification can be included. For this additional step, a filter with a smaller filter may be used. For example, the present inventors have found a pore size of 0.22 pm to be useful at this stage. Filtration with a 0.6/0.2 pm filter may also be used. In some cases, this filtration step may be the final step in the purification process.

Tangential flow filtration (TFF) is a filtration technique in which the fluid to be filtered flows in parallel to a filter rather than perpendicular to it. This arrangement prevents clogging of the filters. TFF is widely used in the bio-pharmaceutical industry, typically for concentration and/or diafiltration. In the present method, TFF is used to concentrate the lentiviral vectors in the media, forming the product. The use of a TFF filter with an appropriate molecular weight cut off also results in the removal of endonuclease and other impurities from the product.

In some cases, a final buffer exchange step can be included so that the final product is in the desired buffer, which may for example be adapted for long term storage or for use in the transduction of cells.

Accordingly, in one embodiment the purification step comprises:

(i) endonuclease treatment;

(ii) filtration through a 0.45 pm filter;

(iii) concentration of the media by tangential flow filtration;

(iv) optional buffer exchange; and

(v) filtration through a 0.6/0.2 pm filter.

The invention will now be further described by way of Examples, which are meant to serve to assist one of ordinary skill in the art in carrying out the invention and are not intended in any way to limit the scope of the invention.

EXAMPLE 1 - VECTOR PRODUCTION

METHODS

The below procedure provides details of the generation of CAT 19 Lenitviral Vector using suspension HEK293T-EX cells from RCB bank. The cells were revived and expanded in Erlenmeyer Shake Flask (SF) and cultivated at 37 °C and 8% CO2 in a FreeStyle™ 293 Expression Medium.

Bioreactor Inoculation and cell expansion prior to transfection.

HEK293T cells grow to maximum cell density (typically to 4xlO ⁶ /mL) in several shake flasks. Cells were harvested and used to inoculate approximately IL in a 10L Flexsafe RM perfusion bag to the final concentration of 5xl0 ⁵ /mL. The 10L bag was mounted on a Xuri Cell Expansion System W25 bioreactor system (Cytiva). The cells were further expanded to 5xlO ⁶ /mL for over 3 days. After inoculation, cell suspensions were cultured in the bioreactor using the following settings: 25 rocks/min, 9° rocking angle, an airflow of 0.2 L/min. Cell expansion in Flexsafe® RM bag prior to transfection was optimized. In run 3 perfusion started 48 Hrs post inculcation at Vi volume per day and increased to 1 volume per day 24 hrs later. In all other runs perfusion started 24 hrs post-inoculation and increased to 1 volume per day 24 hrs later.

Lentivirus Production- Transfection

HEK293T cells adapted to serum-free media were transiently transfected with four plasmids required for third-generation lentivirus (LV) production. Third generation selfinactivating LVs were produced using a four-plasmid system: pVSVf, pGAG-Pol, pRev, and LV transient gene. Plasmids were manufactured by Plasmid Factory or extracted with the EndoFree Plasmid kit (Qiagen) according to the manufacturer’s instructions. Cells were transfected typically on day 3. To facilitate entry of the DNA plasmid into the host cell Polyethyleneimine (PEI) and a total of 34 mg of plasmids were used. The transfection mix was prepared in a safety cabinet according to the protocol (small scale) DNA and PEI were separately mixed with serum-free culture medium, mixtures were combined and incubated at room temperature. The volume corresponding to the transfection mixture was removed from the bioreactor and the mixture was added into the bioreactor. The transfection mix was transferred to IL sterile bottle and welded on to the bioreactor bag. Two different methods of transfecting bioreactors were investigated. The direct assertion of IL transfection mix and pre-diluted transfection mix in an additional IL of FreeStyle™ 293 Expression Medium. Pre dilution of transfection mix in the medium was implemented to improve the mixing of transfection mix in the bioreactor. In runs 3 and 4 perfusion was stopped, and the culture was left overnight at the settings below. In runs, 5, 6 and 7 perfusion was restarted 6 hrs post-transfection at lOL/day flow rate. For determination of transfection efficiency in the bioreactor several ml of the bag content was aspirated 24 hrs post-transfection and the cells were analysed by flow cytometry.

Table 1. Overview of process changes made at the transfection step.

Lentivirus Production- Induction and pH shift strategies

Sodium Butyrate Addition

Cells were induced 24 hrs post-transfection using IM stock concertation of Sodium Butyrate (NaBu). Two different methods of NaBu addition were investigated:

1. The direct addition of 90 ml of NaBu; and

2. Prediluted NaBu in IL of Freestyle medium, to improve the mixing of induction reagent with cell culture.

In runs 5, 6 and 7 perfusion was restarted 6 hrs post-induction at lOL/day flow rate. pH Shift Strategies

Two methods were investigated:

1. pH shift with CO2. After NaBu addition the pH set point was changed to 6.8 and controlled by CO2 only.

2. pH shift with HCL and CO2. 1 : 10 dilution of 33.3% HCL was prepared and filtered through 0.22 pm sterile filter container with approximately 100 ml of HCL.

Solution was welded on to the pump line. The pH setpoint was changed to 6.5 and by using pH control loop and peristaltic pump the appropriate volume of HC1 was pumped into the cell culture. After pH has reached the setpoint the control loop was changed to CO2 control only. Table 2. Overview of process changes made at induction step.

LV production- Vector Particles Harvest

The vector was collected using perfusion mode. The different settings were investigated to improve the vector particle recoveries:

1. 25 rpm 9° angle lOL/day perfusion rate;

2. 15 rpm 12° angle 40L/day perfusion rate;

3. Perfusion in between the harvests

The collection time differed depending on the setting used. 24 hrs for lOL/per day rate and 6 hrs for 40L/per day rate. In runs 3, 4, and 5 when collecting the vector at 40L/per day perfusion rate after 6 hrs perfusion mode was stopped and the settings were changed back to 25 rpm and 9° angle until the next day. In runs 6 and 7 perfusion continued at lOL/day rate for approximately 24hrs. The collection bag was sampled, and the supernatant was centrifuged at 2000g for 5 mins and filtered through a 0.45 pm syringe filter. Samples were stored at -80 °C for further analysis.

Table 3. Overview of process changes at the LV harvest.

RESULTS

Cell counts and cell viability

Assessment of TCC and CV in Flexsafe RM perfusion Bag and CellBag

Figure 1 shows the total cell count and cell culture viability during the expansion phase in Flexfase RM perfusion bag (Sartorius) and CellBag (GE) prior to transfection. Cell expansion in the bioreactor allows decreased inoculation amount, thus decreasing the amount of material and working hours needed to expand the cells before inoculation. Therefore, cell expansion in Flexsafe RM perfusion bag and CellBag prior to Transfection was determined and optimized.

High total cell count was achieved in two production bags, with a target at a minimum of 5xl0 ⁶ viable cells/mL prior to transfection (range: 7.48xl0 ⁶ - 8.76xl0 ⁶ in CellBag and 4.19xl0 ⁶ -5.74xl0 ⁶ in Flexsafe RM Bag) The cell viability in CellBag was significantly lower than in Flexsafe RM Bag (range; 60.6%- 62.33% in CellBag and 87%-94.75% in Flexsafe RM Bag). Therefore, the cells in CellBag were not transfected as the cells did not meet the criteria for transfection (target >90% viable cells/ml).

Flexsafe RM Perfusion Bag TCC during the expansion and LV production phase Figure 2 shows cell growth during expansion and production phases in Flexsafe RM perfusion bag (Sartorius) Each data point is derived from the average value of two measurements taken on the NC-250 Nucleocunter at that timepoint. In run 002 the cells were expanded outside the bioreactor in Enleymer Shake Flasks. In runs Sar004.005.006 and 007 perfusion started 24 hrs post-inoculation and increased to 1 volume per day 24 hrs later. This early perfusion improves cell growth and viability by improving the doubling time to an average of 23hrs., data showed in Table 4. Despite a different time of cell expansion, viable cell counts on the day of transfection didn’t vary significantly between the runs. (range; 4.19xl0 ⁶ - 5.74xl0 ⁶ ). There is no significant difference in cell growth in runs with and without pH shift, (range; 8.36xl0 ⁶ - 1.04xl0 ⁷ at the Harvest II time point) Therefore, HCL addition does not impact cell growth. In contrast, the total cell concertation in runs with no perfusion post-transfection and post-induction showed growth arrest after induction and increased up after perfusion was re-started 24 hrs post-induction.

Table 4: Cell Growth and doubling time during the expansion phase prior to transfection. Sar003 perfusion started 48 hrs post-inoculation and Sar004, 005, 006, 007 perfusion started 24 hrs post-inoculation

Flexsafe RM Perfuison Bag Cell Viability during cell expansion and LV-production. Figure 3 shows cell viability during expansion and production phases in Flexfase RM perfusion bag (Sartorius). Each data point is derived from the average value of two measurements taken on the NC-250 Nucleocunter at that timepoint. Overall increase viability was observed in runs with perfusion posttransfection and post-induction (range; 56-59% no perfusion, 67-69% with perfusion at Harvest II time point) As can be seen from this data cultures with HCL addition compared to those with pH shift with CO2 only didn’t show a significant difference in cell viability at any point during the LV- production (range; 59-66.5% no pH shit and 67-69% with pH shift).

Functional titre

As presented in figure 4A the highest functional titre was obtained already within the first 48 hrs PT (range; 2.24xl0 ⁷ - 2.96xl0 ⁷ TU/ml). On the other hand, samples taken at 96 Hrs and 120 hrs PT time point did not show remarkable productivities (results not shown), and therefore the harvest window was determined to fall between 48 hrs-72 hrs PT. The highest titre at 72 hrs PT time point was obtained in runs with perfusion post transfection and postinduction; 005,006,007 (range; 5.12xl0 ⁶ - 1.43xl0 ⁷ TU/ml).

As shown in Figure 3C this perfusion strategy improves the cell viability and therefore has a positive impact on functional particles concertation. The highest functional titre (1.43xl0 ⁷ TU/ml) for Harvest II (72 hrs PT) was achieved in run 005 when no perfusion between the harvest was applied. Perfusing between the harvest (run 006 and 007) decreases the functional particles concertation in collection bag (006 perfusion bag 3.07xl0 ⁶ TU/ml and 007 3.07xl0 ⁶ TU/ml) probably due to particles escaping to the waste collection bag during overnight perfusion. Figure 4B shows functional particles concertation in the production bag concerning the collection bag..

Physical titre and P:I ratio

P24 protein concertation was analyzed to estimate the total particle concertation at Harvest I and Harvest II as well as in LV collection bags. The physical titre was then used to calculate P:I ratio.

Impact of perfusion membrane on P:I ratio was analyzed and there is no significant difference between production bag and collection bag (Figure 6B). This founding suggests that the setting used during the collection does not have a negative impact on particle functionality. On the other hand the data suggest that the ratio of functional particle to physical particles increases over time, which could be associated with decreased cell culture viability at the corresponding time point (Figure 4A).

Quantification of Trasnfection Effciency of CAR Transgen by Flow Cytometry High transfection efficiency was obtained in both runs (range: 67.9- 76.7% of CAR positive cells). Number of positive cells obtained in bioreactor is highly comparable to the shake flask culture harvested 24 hrs post induction average 61.2 % CAR positive cells.

CONCLUSION

Replacing flask production with a highly integrated bioreactor system reduces labour demands and costs notably. Utilizing the working methods presented here, LV production can be scaled-up to provide a sufficient amount of material for clinical applications.

EXAMPLE 2 - DOWNSTREAM PROCESS

METHODS

Clarification and sterile filtration

Lentiviral material was prepared according to the method described in Example 1 and was kept refrigerated at 2-8 °C until arrangements were made to proceed with downstream processing (DSP).

One day before starting DSP 2.5 units per mL (2.5 U/mL) Denarase® were added to the harvest and the mixture was kept overnight at 2-8°C. An initial filtration step was carried out with a Millipak® 200 0.45 pm filter from Merck Millipore. In-house filter assemblies were created and clarification was performed using a Watson Marlow pump at 50 RPM.

The final filtration step was carried out using an ULTA Capsule HC 0.6/0.2pm 2” filter with 95% volume recovery rate. In-house filter assemblies were created and sterile filtration was performed using Watson Marlow pump at 25 RPM.

Tangential flow filtration (TFF)

TFF was carried out at a shear rate of 4000s' ¹ on an AKTA Flux 6 using a UFP-500-C-6A hollowfibre, with a membrane area of 0.48 m ² . Diafiltration buffer in this experiment was 50 mM HEPES 70 mM NaCl 5% sucrose pH 7.

Analytical

Multiangle dynamic light scattering (MADLS) was performed on the day of DSP execution with fresh samples. Five measurements per sample were taken and analysed according to manufacturer’s instructions (Malvern Panalytical). Samples for other analytical assays were collected and stored at -80°C until the assays could be performed. Statistical data analysis (t-test, One-way Anova) was performed using GraphPad Prism 8.4.2. Recovery calculations after mass balancing were based on the TFF input sample: 100

RESULTS

MADLS

We monitored several particle populations through the purification process (Figure 7) and we were able to calculate vector physical particle recovery (Figure 8):

1.) “Albumin” particle population was not present in any of the samples

2.) “Small aggregates” particle population was also not detected indicating good quality of this harvest.

3.) “Vector” particle population which had hydrodynamic size of 160-174 nm (Figure

7a and b). There was no statistically significant difference in the vector particle population size among different samples in the DSP process according to ordinary one-way Anova. After comparing physical particle concentration for TFF I and TFF- C R we got a >20-fold increase in the concentration which corresponded to >100% recovery of physical particles. The CV% of the TFF I sample was high therefore the highest and lowest values were removed from average particle concentration calculation in order to improve the CV% (33% after removal). This provided us with a 140% recovery of the vector particle population in the TFF concentration step which needs to be interpreted with a degree of caution (Figure 8). This physical particle concentration (pp/mL) and recovery was maintained in diafiltration and sterile filtration.

4.) “Large aggregates” particle population with hydrodynamic size of 413-426 nm was present in all analysed samples except the permeates (Figure 7c and d) and from the SF sample. The removal of aggregates from the SF samples was achieved for the second time. p24 ELISA

After comparing TFF I and harvest samples (H24h and H48h) p24 titre we saw a drop in the titre (Figure 9a). Mass balancing revealed only 68% recovery of p24 in the clarification step, however, this did not match up with the data from previous runs. After comparing the TFF I and TFF-C R there was a 10.2-fold increase in p24 titre which was consistent with 52% recovery of p24 (Figure 9b). The titre was maintained in the diafiltration step and there was small drop of p24 titre in the sterile filtration. The final p24 recovery was 49%. After mass balancing the concentration step we could account for approximately 63% of p24 (Figure 9b) which was very similar to results obtained in previous runs. The recoveries obtained in this run were very similar to p24 results from run II indicating that 50 mM HEPES 70 mM NaCl 5% sucrose pH 7 was performing better as a final formulation compared to Plasma-lyte 148 and TexMACS.

Infectious titre (TU/mL)

We performed the infectivity assay once and we were confident that we seeded the correct number of cells per well was seeded therefore the assay was not repeated (Figure 10a). Clarification recovery was 78%. In TFF concentration step titre was increased 22.3-fold which resulted in 114% recovery of infectious material (Figure 10b). This was maintained in diafiltration and sterile filtration. This was comparable to MADLS results but not to p24 ELISA. As in previous experiments, we did not detect any TU in both permeate samples indicating that all infectious vector was contained in the hollowfibre.

Summary

We obtained >100% recovery of the infectious material (Figure 10b) in the concentration step. Sterile filtration did not cause any loss of infectious material.