Reviewing the process intensification landscape through the introduction of a novel, multitiered classification for downstream processing

A demand for process intensification in biomanufacturing has increased over the past decade due to the ever‐expanding market for biopharmaceuticals. This is largely driven by factors such as a surge in biosimilars as patents expire, an aging population, and a rise in chronic diseases. With these market demands, pressure upon biomanufacturers to produce quality products with rapid turnaround escalates proportionally. Process intensification in biomanufacturing has been well received and accepted across industry based on the demonstration of its benefits of improved productivity and efficiency, while also reducing the cost of goods. However, while these benefits have been shown empirically, the challenges of adopting process intensification into industry remain, from smaller independent start‐up to big pharma. Traditionally, moving from batch to a process intensification scheme has been viewed as an “all or nothing” approach involving continuous bioprocessing, in which the factors of complexity and significant capital costs hinder its adoption. In addition, the literature is crowded with a variety of terms used to describe process intensification (continuous, periodic counter‐current, connected, intensified, steady‐state, etc.). Often, these terms are used inappropriately or as synonyms, which generates confusion in the field. Through a detailed review of current state‐of‐the‐art systems, consumables, and process intensification case studies, we herein propose a defined approach in the implementation of downstream process intensification through a standardized nomenclature and viewing it as distinct independent levels. These can function separately as intensified single‐unit operations or be built upon by integration with other process steps allowing for simple, incremental, cost‐effective implementation of process intensification in the manufacturing of biopharmaceuticals.

Global biopharmaceuticals marketing reports have reported a market size of 407.72 billion (USD) in 2023 which is expected to increase at a compound annual growth rate (CAGR) of 8.03% by 2028, resulting in a market size of 651.78 billion (USD) (Modor Intelligence, 2023).This growth is driven by the demand of treatments toward a variety of complex diseases with innovative methodologies and modalities.For biomanufacturers to be able to meet these requests and remain competitive, there is a need to increase speed to market, efficiency, and productivity, all while reducing the overall cost of goods.To address this, the biopharma industry has steered toward process intensification.
While the goals of process intensification are well defined, there currently is not a unified definition of process intensification within the biomanufacturing industry.Along with the various definitions, a myriad of terms has also been associated with it, which further increases confusion.The terms of continuous, end-to-end, integrated, connected, closed, and hybrid are commonly seen and often used interchangeably (Chen et al., 2022;Coffman et al., 2021;De Luca et al., 2020;Godawat et al., 2015;Goudar et al., 2015;Karst et al., 2018;Yang et al., 2019).Yet, regardless of the number of terms used and definitions applied, the goal is always that of increasing productivity and operational efficiency.
With the rising biopharmaceutical market demand, an elevation in the cost of development and manufacturing of pharmaceuticals is also observed.This allows us to clearly outline the objectives and inherent benefits of process intensification as detailed in Figure 1.Through the implementation of one or more of these factors, two primary goals can be achieved.The first is the key business driver: a reduction in the overall cost of goods (Amasawa et al., 2021;Gupta et al., 2021;Hummel et al., 2019;Pollard et al., 2016).Secondly, a more sustainable process can be achieved (Kumar et al., 2020).
A comparison is often made to the automotive industry, where automation, continuous manufacturing, and lean practices are commonplace.Regarding the field of autonomous driving, a well-defined structure involving six levels has been established.The classification of these levels begins at level 0, in which an automobile has no freestanding driving capabilities, and through a series of incremental advancements, the levels increase to level 5, in which an automobile has full autonomy and is able to operate independently under any road or traffic conditions (Xu et al., 2022).Within levels 1-4 are varying degrees of automation of the vehicle.More so, these levels have been defined by the Society of Automotive Engineers (SAE) and adopted by the Department of Transportation, meaning that throughout the industry, there is alignment and harmonization (US Department of Transportation, 2018).
In comparison, there is also a desire and need for the biopharma industry to eventually reach this same level of process intensification, standardization, and harmonization, despite the challenges.However, process intensification should not be seen as an "all or nothing" approach, and while fully automated continuous manufacturing may be seen as the ideal future state, the real need may lie somewhere between a batch process and a fully intensified continuous process.
Based on the analysis conducted herein, we have observed a number of different terminologies used to describe or explain process intensification, including but not limited to continuous processing, pseudocontinuous processing, and integrated processing.These concepts are clearly defined relative to their differences from traditional batch processing, yet the distinction in these terminologies becomes less clear when compared to each other.These terminologies provide a vague description of what process intensification is without clearly defining the differences between these various intensification methodologies.While the specific definition of process intensification varies from one methodology to another, these definitions are typically underpinned by several key factors, as mentioned earlier: increasing efficiency, reducing footprint, increasing productivity, and so forth, all while reducing the overall cost of goods.
F I G U R E 1 Overview representing the benefits of biomanufacturing process intensification.Benefits outlined with a focus on cost, time savings, and quality.E2E, End to end; STY, space time yield (volumetric measurement of productivity in units of grams/liter/hour).
By reviewing this literature, we have categorized process intensification into defined levels based on specific criteria (e.g., standalone unit operations, connected unit operations, automation, closed processing, steady-state flow operation, etc.), with each level increasing in system and technology complexity as shown in Figure 2.
Designating such a level-based approach can facilitate greater standardization and harmonization across industries while also allowing for easier alignment to regulatory guidance for commercialization of therapies produced using heightened methodologies.

| THE BENEFITS OF PROCESS INTENSIFICATION
The benefits of process intensification have been well documented by comparing batch processing to various forms of process intensification and presenting the product quality, process productivity, and economic benefits (Arnold et al., 2019;Gupta et al., 2021;Müller et al., 2022;Schaber et al., 2011;Walther et al., 2015).One key initial proof of concept work began with a monoclonal antibody-based process (Warikoo et al., 2012), in which a perfusion-based bioreactor was connected to a periodic counter-current chromatography system for both monoclonal antibodies and recombinant enzyme processes.The system was able to run without decay for a period of 30 days and demonstrated a reduction in overall footprint, reduction of nonvalue-added steps, and an increase in resin utilization, all while maintaining product quality.Since then, other groups have included a greater number of unit operations as both proof of concept and development toward commercialization.Examples include process intensification of the capture step by multicolumn chromatography (Coolbaugh et al., 2021;Kateja et al., 2021;Shi et al., 2021), the inclusion of a multicolumn chromatography capture step followed by connected polishing steps with ion exchange membrane adsorbers (Gjoka et al., 2017), continuous viral inactivation (Gillespie et al., 2019;Martins et al., 2020) and single-pass TFF (Clutterbuck et al., 2017;Ding et al., 2022;Elich et al., 2019).Processes have more recently developed even further towards that of a fully integrated continuous downstream approach (Coolbaugh et al., 2021;Ramos et al., 2023).While these recent processes are not fully optimized, they do provide an excellent framework and proof of concept that has brought us ever closer to the end goal of true integrated continuous biomanufacturing.
Certain critical commercial advancements have also been made towards the implementation of process intensification within the biomanufacturing space.This has been observed globally within the United States, Europe, Japan, and South Korea and includes such companies as Vertex, Eli Lily, Pfizer, Novartis, Johnson & Johnson, Shionogi, and SK Biotech. Since 2017, Vertex, Pfizer, Johnson & Johnson, and Eli Lilly have all gained FDA approval of small molecule-based pharmaceuticals in which the manufacturing process contains an aspect of process intensification within the process (Badman et al., 2019).
Notably, there has been recent work carried out by NIIMBL, where an attempt to provide guidance to the different options of intensification has been presented (National Institute for Innovation in Manufacturing Biopharmaceuticals, 2022).This case study illustrates two major options, an integrated continuous process and a periodic batch process.In our F I G U R E 2 Overview of process intensification levels 0-3.1.Level 0 representing the traditional fed-batch operation.Level 1 involves the intensification of a single-unit operation within a DSP process.Within Level 2, more than one unit operation has connectivity, both via systems and software orchestration.Level 3 and 3.1 both represent a fully connected and closed end-to-end system where level 3.0 maintains DSP bind and elute steps and 3.1 is a continuous flow-through process.Both levels 3.0 and 3.1 present the highest level of complexity in steady-state flow and automation control via software orchestration.| 879 view, this is the first paper to show clustering and options, although not fully exhaustive.While these examples have cited the use of process intensification via a continuous process, the actual applications have largely been a hybrid-based approach between batch and continuous processing.The implementation of downstream process intensification has been piecemeal, with a focus on individual steps, primarily that of the capture step, for the purpose of integration and connection with the USP perfusion bioreactors.To date, the process steps of viral inactivation, viral filtration, and final ultrafiltration/diafiltration still fall under batch processing for biopharmaceuticals in which continuous operation steps are implemented (Coffman et al., 2021).
In 2019, BiosanaPharma was noted as being the first to achieve biomanufacturing of a pharmaceutical utilizing a fully continuous connected biomanufacturing process towards the production of biosimilar omalizumab (Xolair ® ).Though the details of their "3C Platform" process have not been disclosed at this time, the biosimilar produced from the process received approval from the Australian Bellberry Human Research Ethics Committee to begin Phase I clinical trials (Vuksanaj, 2019).Biosana's press release of the Phase I clinical trials showed that through the 3C continuous manufacturing process, the biosimilar maintained the same level of product quality, safety, and efficacy as the originator.More so, manufacturing production costs were reduced by 90% (BiosanaPharma, 2021).
In addition to the benefits relating to the manufacturing process, others have cited secondary benefits from process intensification.These include an increase in national security through the reduction in dependency of foreign suppliers, the ability to respond to drug shortages in a more rapid manner, certain tax incentives for investment or production utilizing integrated continuous manufacturing, and regulatory incentives that could provide an accelerated approval process (Badman et al., 2019).Personnel-based cost benefits also begin to emerge.In an automated connected process, nonvalue-adding tasks of monitoring and controlling a process are removed.The cost savings relating to labor through the implantation of process intensification has been demonstrated by the Novartis-MIT Center for Continuous Manufacturing (Schaber et al., 2011).In conjunction, when performing a risk factor assessment, having an automated steady state process also reduces the risk factor that unfortunately is associated with human error.Deviations occurring within a manufacturing process require thorough investigations which may result in a pause in operations and require efforts of resources to document and resolve the issue properly.Therefore, it can then be said that each deviation is associated with a significant cost.Through a higher level of process intensification involving automation, these risks may be better mitigated.

| THE CHALLENGES OF PROCESS INTENSIFICATION
Despite these detailed process benefits along with advancements through systems and consumables in downstream biomanufacturing, overall implementation of process intensification into industry has not been adopted as quickly as hoped (Erickson et al., 2021;Fisher et al., 2019;Kumar et al., 2020).The reasonings for this have been published in detail from both a business and technical point of view (Rathore et al., 2023).First, there is an engrained mentality of process intensification processing being too complex.Fortunately, this perception is beginning to change as both end users and vendors are working together to develop more user-friendly systems and training programs.
Second, from a business standpoint, a critical financial assessment must be considered towards the company goals beyond that of cost of goods alone (Monge & Sinclair, 2011).To implement process intensification, a changeover must occur to implement within an existing biomanufacturing facility which requires capital expenditure (CAPEX) for equipment and facility design or layout, training of individuals, new documentation, equipment integration with existing systems, and potential downtime.Technical issues relating to long system run times and the ability to maintain adequate control of biocontamination sterility throughout a process conflict with flexibility and the need for rapid changeover (Croughan et al., 2015;Konstantinov & Cooney, 2015).
Finally, regulatory guidelines have long been considered an additional key facet hindering the integration of process intensification (Baxendale et al., 2015;Vanhoorne & Vervaet, 2020).Without regulatory guidelines, the global harmonization of process intensification schemes is not possible.Fortunately, at the time of this writing, the FDA has since released the new ICH Q13 guidelines for the Continuous Manufacturing of Drug Substances and Drug Products (Federal Food and Drug Administration, 2023).Within the scope are guidelines for both a hybrid-based continuous process consisting of a combination of batch and integrated steps, as well as guidelines for a fully continuous process.Also contained within this document are pivotal considerations in process design, control strategy, process validation and regulatory consideration.With the advent of these new guidelines, regulatory agencies, and industry now have appropriate guidance on the implementation of continuous manufacturing globally under cGMP conditions for a variety of modalities, covering both that of large and small molecules.This is an important aspect as the complexities of various diseases are now being evaluated for treatment against a wider range of modalities than ever before.

| CHOOSING A PROCESS INTENSIFICATION SCHEME FOR A PROCESS
In designing an intensified process for a biopharmaceutical, several considerations must be taken into account as well as an understanding that there is not a "one size fits all" process approach.This can be dependent on the molecule of interest or the size of the facility, to cite only two examples.More so, one must prioritize process intensification goals with the knowledge and acceptance that not all process intensification benefits will be achieved with maximum optimization.As an example, if an overall goal is to decrease the overall processing or cycle time of the chromatographic capture step in the process, this may, in turn, reduce the dynamic binding capacity of the chromatographic medium, resulting in a decreased mass product output and thus the need for an increase in the number of total cycles to be executed to obtain the final desired product mass.Figure 3 shows a means to visualize these options to consider.Each component of this toolbox influences each other in both a forward and backward direction.The modality drives which consumables can be best utilized, which in turn may influence the system or vice versa.Upon defining the modality and prioritizing the goals one wishes to achieve, the subsequent decision factors are then those relating to systems, consumables, hardware, and software.
Automation/control must be considered if a main driver is to develop a fully continuous process.As mentioned, many options and configurations have been developed in recent years towards achieving process intensification goals that provide alternatives to batch processes.

| Level 0
Within the proposed framework, Level 0 would be defined as standard batch processing, which can be considered as the baseline processing method for modern biologics (Croughan et al., 2015;Kumar et al., 2020).
Due to a long history of developing and producing medicinal products in this manner, batch processing is well-understood and well-serviced by the biopharma market-the majority of systems and technologies currently in the market (for both upstream and downstream capabilities) are built around batch implementation, driving product quality through competition, as the industry is not wholly reliant on one supplier for standard upstream or downstream technologies.Processing at Level 0 provides the greatest level of fluidity and freedom in process design.As each unit operation is standalone, changes to one part of the process are easier to manage compared to more intensified processes that rely on greater interconnection between unit operations.Batch unit operations can be developed, validated, and characterized in isolation and owing to its ubiquitous use within industry, is well advised by regulatory guidance (US Food and Drug Administration, 1996, 2001, 2020).
The limitations of Level 0 help to define the structure of our proposed leveling system for intensification.Batch processing, by nature, puts a ceiling on the level of productivity that can be achieved (Singh et al., 2012) therefore limiting the amount of product that can be made, necessitating scale-up to extremely large volumes (2,000 L + ), requiring significant facility investment along with all the issues associated with high-volume liquid handling (Hernandez, 2016).At a point, volumetric scale-up becomes extremely problematic as certain technologies (e.g., filtration devices) may not scale linearly beyond a certain point, resulting in a cap on volumetric scalability.The bigger and more productive USP process becomes, the closer to this volumetric cap the DSP process will be, potentially necessitating the implementation of costly scale-out solutions, more so compared to high-productivity processes.This idea of a productivity cap underpins the entire need for the various methodologies of process intensification.
As reported in a recent survey, the leading factors hindering continuous adoption, again, relate to cost, process controls, and regulatory uncertainty.Yet, 79% of those surveyed within this report agreed that we will see the implementation of continuous biomanufacturing at the commercial scale, while only 30% agreed that F I G U R E 3 Overview of process intensification factors towards the selection of a downstream process involving process intensification.Examples listed for overall process intensification goals, systems, and consumables.ADC, antibody drug conjugate; AAV, adeno-associated virus; AV, adenovirus; COG, cost of goods; HIC, hydrophobic interaction chromatography; IEX, ion exchange chromatography; LV, lentivirus; mAb, monoclonal antibodies; r. proteins, recombinant proteins.continuous is simply not needed.More so, of those surveyed, 45.1% stated that downstream continuous operations are currently in use at their facilities in 2023, which is an increase of 12% in comparison to that of 2022 (Schmidt & Winterhalter, 2023).Despite the above-mentioned limitations and statistics provided of continuous downstream adoption, batch processing remains as the leading biomanufacturing processing operation mode for modern and advanced therapeutics.It is unlikely that this paradigm will shift anytime soon-the market has developed around batch processing, combined with its ease of implementation making batch processing a natural starting point for the development of future biologics.
Nevertheless, it is important to consider that varying levels of intensification do currently exist for modalities where the productivity cap has reached a plateau or scale-up is not possible due to bioreactor and/or facility constraints (e.g., monoclonal antibodies) (Nitika et al., 2023) along with other modalities where new technologies can help improve process efficiency (Adamson-Small et al., 2020).

| Level 1
Level 1 of the proposed process intensification framework concerns single-step intensification, either by systems or the adoption of novel methodologies or technologies that allow for a greater level of process productivity to be achieved.This includes multicolumn chromatography (MCC) systems such as the BioSMB and BioSC (Sartorius), AKTA PCC (Cytiva), Octave (Tosoh), and Contichrom (YMC) which have all been developed towards continuous chromatography (Carvalho & Castilho, 2017).Consumable technologies are also making an impact toward process intensification, such as convective membrane adsorbers, monolithic chromatography columns, Single-Pass Tangential Flow Filtration (SPTFF), and nanofibers.
Recent advances from Sartorius and Gore in membrane technology and Cytiva using fiber-based technology have produced matrices functionalized with Protein A ligands that may now be used towards the monoclonal antibody processes capture step combined with a strategy termed Rapid Cycling Chromatography (RCC) (Jungbauer, 2013).
As Level 1 intensification operates in isolation, that is, on a single step within a process, the rest of the downstream process may not need to change much, if at all, to accommodate.This makes Level 1 intensification a good next step from Level 0, serving as an example for how process intensification can benefit DSP without necessitating a complete redevelopment of the downstream process.However, in addition to the direct benefits of PI, it is important to understand the indirect benefits of Level 1 intensification, where the benefits from the intensified step (e.g., increased purity, increased concentration, etc.,) may be realized in subsequent DSP steps (increased filter throughput, smaller process volumes), allowing further optimization of these other, non-intensified steps as a result.
There are several main drivers for adopting Level 1 intensified chromatography technologies.Firstly, by changing the mode of operation to RCC or MCC, greater stationary phase utilization can be achieved (Chen et al., 2021), allowing for smaller column volumes, which is important in the context of the price of chromatography media (Bansode et al., 2022).Smaller column volumes provide several advantages, notably, easier column handling, lower buffer consumption, lower processing volumes, and reduction in footprint for buffer tanks and systems.
Level 1 intensification is geared heavily towards a focus on chromatography.In general, chromatography steps are often seen as the main candidate for level 1 process intensification as they are usually the rate-limiting steps in terms of productivity, and as discussed previously, chromatography sorbents are often expensive.
Intensification of TFF via SPTFF is an attractive Level 1 target, as it enables later levels of intensification (Levels 2-3.1), but also helps to eliminate some of the problems associated with TFF, such as processing time, shear stress, and footprint (Madsen et al., 2022).
While there are instances in which the situation encourages TFF to be performed in a batch-wise manner (Coffman et al., 2021), TFF generally is a major obstacle to continuous processing due to how the product is handled-multiple cycles of the product through the membrane prevents steady-state flow, due to the batch operation of traditional TFF.SPTFF alleviates this issue by removing the recirculation pathway in the retentate, allowing for a constant output of product from the SPTFF device.
With the increasing popularity of novel technologies such as convective chromatography media, intensified chromatography devices are becoming more commonplace within the industry, especially for advanced therapies, such as the use of membranebased devices for AAV (Hejmowski et al., 2022) and monolith-based devices for mRNA (Luisetto et al., 2022).In addition, for proteinbased therapeutics, such as monoclonal antibodies, suppliers are looking to push beyond the limitations of resin-based chromatography to convective devices (Busse et al., 2022;Cytiva, 2020;Gore, 2022) to overcome the aforementioned limitations of resin as a capturing step.Alongside these new protein A technologies, RCC and MCC have shown great promise in the purification of mAb products, utilizing either resin or next-generation protein A devices (Angarita et al., 2015;Brämer et al., 2019).Lastly, dedicated systems for Level 1 process intensification are beginning to gain traction, both due to the diversity of systems offered and user familiarity to concepts such as continuous chromatography, simulated moving bed, and so forth (Baur et al., 2016;Girard et al., 2015).While these sorts of systems have been typically limited to protein-based therapies, recent work has been done using intensified systems for advanced therapies (Mendes, Bergman, et al., 2022).However, most of the work published in the literature today has been done by academic groups, and the translation into GMP processes is still to be seen.In addition to chromatography, Tangential Flow Depth Filtration (TFDF) by Repligen has been demonstrated for the clarification of lentiviral and adeno-associated vectors as a way to combine the benefits of TFF and depth filtration into a single step (Mendes, Fernandes, et al., 2022;Williams et al., 2020).
So where are the challenges?In most cases, to get from Level 0 to Level 1 will require re-optimization of the step, requiring additional expertise and increased CAPEX.There are added layers of complexity compared to Level 0 when exploring options such as Simulated Moving Bed (SMB) or RCC, requiring investment into training or acquisition of experienced staff.This, alongside the potential purchase of new systems, could represent a significant financial commitment that Level 0 would not have.That being said, some strategies, such as parallel batching, could be easy to implement as minimal changes to the process parameters are needed.For Level 0 processes that are already in commercial manufacturing, adoption of Level 1 intensification strategies would require revalidation of the process, especially if the process is in late phase III or commercial due to the potential impact on product quality stemming from changes from either the critical process parameters (CPPs) or critical material attributes (CMAs).It is important to note that should these changes be implemented after confirmatory trials or after approval of the drug; then a comparability exercise would be required, including physiochemical and biological in vitro studies, as well as potentially clinical pharmacokinetic and pharmacodynamic comparability studies (European Medicines Agency, 2007;Federal Food and Drug Administration,18, 2018).For implementing process intensification into an already existing commercial product, this may be the biggest hurdle to overcome (Drobnjakovic et al., 2023;Nasr et al., 2017).
However, new second-generation processes are being developed and implemented, which could circumvent this issue.One of the more significant examples that could be followed by others in the biomanufacturing industry comes from Amgen.In this example, Amgen introduced process intensification through continuous operations of existing batch processes through the build-out of new facilities in Tuas, Singapore (Hernandez, 2017) and Rhode Island, USA (Enright, 2020).Both of these facilities are a fraction of the square footage size, use less raw materials, and produce less waste, yet able to maintain the same level of productivity while reducing costs and increasing speed to market.Nevertheless, Level 1 intensification is likely to become more standard for chromatography as time passes, especially with consumables like convective membrane chromatography devices.
Implementation of Level 1 intensification through systems or methodology should become more common for protein-based therapeutics as time goes on due to the productivity cap associated with batch chromatography using resin-based consumables (Brinkmann & Elouafiq, 2021).For filtration-based steps to be included as a standard in intensification, it will be influenced in conjunction with the adoption of Level 1 chromatography steps, as developers seek to find ways to expand intensification from just one unit operation (Level 1) to multiple, connected unit operations (Level 2).

| Level 2
While Level 0 and Level 1 are focused on addressing intensification of single unit operations, Level 2 begins looking holistically at the process with the connection of two or more unit operations running simultaneously.This may be referred to as a clustered or connected intensified process.Level 2 intensification allows for process time compression, as the previous unit operation does not need to finish for the next one to start, unlike Level 0 or Level 1 processes.
Intensification in this manner can be achieved through the use of systems (e.g., Fujifilm's SymphonX, PAK BioSolutions and Sartorius' BioSC) or by making changes to the operating parameters of multiple Level 0/1 unit operations, using surge tanks where needed (Coolbaugh et al., 2021).In general, Level 2 intensification centers around at least one chromatography step due to the aforementioned issues with bind and elute chromatography being a rate-limiting step.
Level 2 can be considered the first step towards a fully continuous process-the ultimate end goal for process intensification in the biopharma industry (Croughan et al., 2015).As mentioned above, the ability to run multiple unit operations simultaneously contributes to considerable time savings within a process and whilst there are systems available that can run multiple unit operations at once, Level 2 intensification can be implemented with standard batch unit operations (under the right conditions), without the need for new systems or technologies.Therefore, in some cases, Level 2 intensified processes may require less re-designing of the process compared to some Level 1 options, thereby making it attractive from a CAPEX point of view.In addition, where systems that can run multiple unit operations are involved, the benefits here are realized in footprint and CAPEX savings, due to the reduction in the number of systems needed.Lastly, implementing Level 2 intensification may achieve the desired productivity gains from the connection of 2 to 3-unit operations without the complexity of needing to adopt a fully continuous process.
Despite the appeal of and drive towards continuous processing, the adoption of Level 2 processing does have its own limitations that need to be carefully considered and worked around for successful implementation.Combining two steps to run alongside one another requires either the use of surge tanks, due to imperfection in flow rates or both steps to be running at similar flow rates.Due to this, one step may need to be run sub-optimally to facilitate connection to the other step e.g., reduction in clarification filtration flux to maintain an optimal residence time at chromatography loading.Understanding of design space, CQAs and CPPs is critical for Level 2 to understand what effect any changes may have on the process after implementation.Greater emphasis on the understanding of the process design space is needed for Level 2 intensification compared to Level 0/1.In addition, Level 2 intensification may require the need for automation software enabling unit orchestration, depending on which unit operations are connected.This could be considered a limiting factor at this level, given that there are other ways to intensify the process to Level 2 without introducing the complexity of automation orchestration (Whitford, 2022).However, within Level 2 intensification, users have the choice to either use various, nearly unlimited unit operations connected by a software orchestration layer or move towards orchestration of unit operations on a system level, with reduced options of available systems and unit operations.
One Level 2 process described in the literature is the Accelerated Seamless Antibody Purification (ASAP) process from Sanofi (Mothes et al., 2017).This process well demonstrates a Level 2 Process Intensification scheme through the interconnectivity of at least two process steps.In this example, convective media membrane adsorbers of Protein A capture and ion-exchange polishing chemistries are utilized within a monoclonal antibody process.Connectivity of process steps is achieved primarily through the cation exchange to anion exchange steps, while the protein A capture eluate may be processed via in-line viral inactivation or directly loaded on to a subsequent column.The benefits of this implementation of process intensification level 2 are highlighted in processing time, reduction of nonvalue-based tasks, productivity, and reduced capital costs.
Nevertheless, with the constant development of innovative technologies and a concerted effort within the industry to push towards more productive downstream processes, Level 2 intensification can rightfully be considered for many modern therapeutics.This is especially true for processes that rely on convective chromatography media, where the faster flow rates used for these devices facilitates connection with filtration steps.In addition, multi-step systems can really help to push the envelope with regard to intensification of resin-based processes.

| Level 3
While Level 2 can be very much considered an introduction towards continuous processing, Level 3 intensification concerns fully continuous processing, involving the complete integration of all downstream unit operations (chromatography, TFF, viral filtration, sterile filtration) with a steady-state flow, including automation control systems and minimal surge tanks.Level 3 intensification allows DSP to fully meet the requirements of a perfusion-based USP platform, resulting in longer processing times, as the process can now continuously purify the material generated from the bioreactor to drug substance.In the context of commercialized processes, Level 3 intensification could allow for RTRT (Real Time Release Testing) of the drug product in processes where PAT (Process Analytical Technologies) have been implemented.This would represent a significant time saving, given that the release of the drug product in a batch process can often be several months after the conclusion of a batch (Siemens, 2015).
The transition from Level 2 to a fully continuous process requires a greater leap forward compared to Level 0 to Level 1 or 2, in part due to software orchestration, which is mandatory for the running of Level 3 intensified processes.While certain aspects of software orchestration can be leveraged from level 2 to level 3, for a fully continuous system, each system must be able to communicate.In cases of differing vendors for certain systems or automation control this presents a challenge.A common example of noncommunication would be between that of systems controlled by Delta V versus Siemens.Software orchestration allows for increased consistency in the drug product that can be generated on a daily basis with little impact of variation towards CQAs, by responding to process changes in real time (Mahal et al., 2021).However, because every unit operation is performing in parallel, greater time savings compared to Level 0, 1, or 2 processes can be achieved.As intermediate processing volumes for Level 3 processes tend to be much smaller, the footprint needed for these processes is also reduced, resulting in greater cleanroom utilization.This is further amplified by the nature of continuous processing, meaning there is little to no downtime in the facility compared to traditional batch processes.
In general, Level 3 intensification remains challenging for suppliers and manufacturers.However, there are several examples of its use in industry which helps set the standard for others to follow (Klutz et al., 2015).That being said, the complexity in combining system technologies with automation technology, such as Supervisory Control and Data Acquisition (SCADA) systems or PAT methodologies are a roadblock to many prospective companies, as the scope of process development extends beyond purely process technology and into digital technology, requiring extra skillsets which are not readily available at every company.It must also be considered that the high financial barrier to entry for Level 3 intensified processing, both in terms of CAPEX and FTE required, might mean that this level is limited to companies with enough resource to address these issues, at least until this barrier to entry is lowered.In addition, direct flow filtration is difficult to implement into a continuous process due to the limited lifetime as a consequence of filter fouling.The current strategy involves swapping filters out mid campaign once their lifetime has been reached (Thakur et al., 2020), introducing risks into the process, such as bioburden or other adventitious agents.This is further complicated by the need for closed processing at Level 3 intensification to maintain adequate control of biocontamination over the long campaign times-single-use, aseptic technologies can help mitigate this risk.The recently published ICH Q5A (R2) guidelines on viral safety for continuous manufacturing help provide users with guidelines on how to implement viral inactivation and viral filtration in a continuous manner within the process (European Medicines Agency, 2022).
While the concept and implementation of continuous processing is constantly evolving, its main application in modern therapeutics seems localized to protein-based therapeutics, such as the MoBiDiK (David et al., 2020;Klutz, Holtmann, et al., 2016;Klutz, Lobedann, et al., 2016) and BioContinuum processes mentioned earlier, thereby limiting its implementation to a percentage of pharma companies.
One must also consider that regulatory guidance for continuous processing is not yet at the level of batch processing, leading to elements of uncertainty and risk in these processes.As mentioned previously, the FDA have published guidelines for industry on continuous manufacturing of drug substances and drug products which will help to standardize approaches to market for therapeutics manufactured using continuous technologies and tangentially, mitigate any perceived regulatory risk.Long-term, the industry seems keen to consciously move towards Level 3 intensification as companies seek productivity gains that surpass what is possible with standard batch (Level 0) processes (Croughan et al., 2015).

| Level 3.1
Advancements in mAb purification platforms have continuously been evolving over the years.This has been observed from traditional bind and elute steps for all chromatographic steps to a now commonplace incorporation of anion exchange flow-through step, upwards to that of a single bind and elute capture followed by both intermediate polishing steps being operated in flow through mode (Coolbaugh et al., 2021;Mothes et al., 2017).This evolution in chromatography operations is also considered for the definition of process intensification levels.
Level 3.1 is labeled as such to be seen more as a modification to Level 3 as opposed to providing something entirely different altogether.The key difference between Level 3 and 3.1 is that the molecule does not enter a stationary phase within the process.This means replacing any bind and elute chromatography steps with flowthrough steps and the removal or modification of steps that rely on recirculation (e.g., TFF).Level 3.1 intensification should be considered to be squeezing the last drops of productivity out of Level 3 processes.
Through the removal of bind and elute chromatography steps, the process allows for the decrease or complete elimination of wash and elution steps compared to that of bind and elute steps.This, in turn, provides a more economical and greener process as the ratio of required materials to produce an equal amount of product is decreased, also known as the process mass intensity (PMI).Here, convective chromatography technology could be extremely attractive due to the increased flow rates they can achieve compared to resinbased chromatography, reducing the overall process time even further.Additionally, as all chromatography steps in Level 3.1 would be operated in flow-through mode, the downside of poor resolution from membranes becomes less of a factor when taking overall benefits into account.In contrast, as the molecule does not stop, there is an increased layer of complexity and difficulty in designing a Level 3.1 process compared to Level 3 processes, with regard to the design of the process in replacing the initial capture bind and elute step.
With Level 3.1 processes employing flow-through chromatography, it is important to note that this results in the absence of any affinity capture steps, which is where intensification efforts of Levels 1 and 2 are focused.While affinity chromatography is generally preferred for any product it is available for, its limitations are well documented: high cost, slow, and product-specific (Pina et al., 2014;Ramos-de-la-Peña et al., 2019).Overcoming these limitations with a non-affinity capture step is a key driver for the adoption of Level 3.1.
Capto Core from Cytiva for the capture of Lentiviral vectors is an example of where flow-through technology-has been employed towards the capture of a target molecule (Leung et al., 2020).
While level 3.1 processes for modern therapeutics are still some ways away from being realized in a commercial setting, the initial efforts have begun.Work has been done virtually to demonstrate the advantages in COGs reduction by adopting this level of intensification towards a monoclonal antibody process, with a comprehensive flow through simulation at 1000 L manufacturing scale compared to a standard bind and elute capture process (Prouzeau et al., 2021).
While Protein A has demonstrated itself well towards the development of a platform-based process for mAbs for hundreds of processes and products, it is difficult to foresee a monoclonal antibody process without it.Yet, as previously mentioned for affinitybased media, it is not without its drawbacks of high cost, caustic stability, and low pH elution conditions.Researchers have begun to move towards its removal bringing about a complete flow-through process consisting of depth filtration towards impurity removal versus Protein A bind and elute (Yamada et al., 2017) as well as a process involving of the replacement of Protein A with an activated carbon filter membrane followed by an anion exchange membrane for a column-free process (Ishihara et al., 2019).In both processes, the impurity and recovery levels were comparable with that of a bind and elute Protein A-based process.
Through another approach, counter-current tangential chromatography (CCTC) has shown promise both academically and in a commercial setting with Chromatan (Dutton, 2015;Fedorenko et al., 2020;Mohammadzadehmarandi et al., 2023).What is advantageous here is that CCTC achieves high-resolution separation in a column-free manner, increases productivity with reduced buffer consumption, and operates in a true steady state.This technology brings us yet even closer to that of a truly continuous process.

| Scenarios for process intensification
Taking the levels of process intensification described herein, an example output is provided as shown in Table 1 and Figure 4.With a focus on a 500 kg mAb production for an annual campaign per year, the Protein A capture step was analyzed against that of traditional batch operation utilizing six potential process intensification scenarios established and evaluated.These scenarios provide a single-step overview on the impact of process intensification for that of the capture step, as the monoclonal antibody Protein A capture step has the highest impact on the cost of goods (Bansode et al., 2022;Tosoh, 2018).As shown in Table 1, downstream process intensification steps of capture included sequential multicolumn capture chromatography (S-MCC), parallel batch multicolumn chromatography (PB-MCC) operation, rapid cycling chromatography (RCC) with convective diffusive membrane media, sequential multi-membrane chromatography (S-MMC) and operation in both a bind and elute (B&E) mode versus that of flow-through (FT) for polishing steps.to a second medium, thereby utilizing its full capacity.Alternatively, parallel batch multi-column is operated as in batch chromatography, with discontinuous loading on a system capable of running multiple columns (Cytiva PCC, Sartorius BioSMB).Therefore, multiple runs can be performed in an automated manner.The output, shown in Figure 4, was determined for resin utilization, buffer PMI, overall productivity, and operating expenditure (OPEX) cost involving resin and buffer consumption.By taking all four charts into account, the fact that there must be a prioritization of biomanufacturing goals (i.e., productivity, resin utilization, buffer consumption, etc.) is further reinforced that moving towards a fully continuous connected end-toend process does not equate to that of maximal optimization for all parameters.If viewing connected end to end processes of D1, D2, and E in comparison with that of hybrid or intensified unit operations within Figure 4, we see that being 100% optimal for all four categories cannot be obtained.Therefore, it is the levels of process intensification that should be considered, and it may, in fact, be only a level 1 intensified process for a capture step that allows one to achieve their goal.If the goal is greater, i.e., further reduction in footprint, total campaign time, and so forth, then this initial evaluation would need to be expanded on beyond that of the capture step, and analysis would need to be performed on the process as a whole, taking mass balance into account for each process step and the different level of intensification from level 1 to 3. Models for this have been performed, demonstrating the benefits and challenges of process intensification upon individual unit operations and a fully continuous end to end process (Lin et al., 2021;Xenopoulos, 2015).Overall, choosing the appropriate level of process intensification involves numerous considerations that need to be accounted for and will vary based on the goals that a process needs to achieve.

| FUTURE STATE: CHALLENGES OF APPLYING PI TO NEWER MODALITIES OR THERAPIES
Whereas intensification and continuous processing of protein-based therapeutics is largely understood, with a clear roadmap on how to get there, intensification of newer modalities (gene therapy vectors, mRNA) still has its own challenges and hurdles to overcome namely (i) stability and (ii) overall process recovery.While expected process yields of batch mAb processes are typically in the 80%-95% range (Gillespie et al., 2014), many newer therapeutics, especially viral vectors, are attaining process yields of approximately 30% (Lyle et al., 2023).Intensification of low titer processes is not really a priority, rather the current focus here is clearly on developing processes that can achieve better yields, whilst maintaining the Illustrating the overall benefit or drawback of DSP Process Intensification scenarios against that of traditional batch processing outlined in Table 1.(a) Overall resin utilization against that of g of mAb produced (g/L), (b) Buffer process mass intensity (PMI) measuring liters of buffer against kg of mAb produced, (c) Overall productivity of each scenario measured in grams of product per liter per hour (g/L/h), and (d) the operating expenditure for each scenario.
stability of the product, especially for sensitive products such as Lentivirus or RNA-based therapies.However, that does not mean that recent technologies cannot help to bridge this gap, whilst also contributing to process intensification.As mentioned earlier, TFDF technology from Repligen is seeing promising results in AAV and LV processes.Convective chromatography also sees much success with new modalities, although this is largely owing to their ability to bind large molecules, such as viruses (Žigon et al., 2022) Alternatively, with AAV, the recent developments in viral vector affinity resin technologies coupled with multi-column chromatography is allowing for greater implementation of process intensification (Mendes, Bergman, et al., 2022) that could potentially extend across all serotypes in the future.When considering intensification for new modalities, the real consideration is the end-user desire.Looking at these newer modalities would suggest that intensification beyond Level 1 may never be needed.Consider mRNA, where the field is moving towards smaller commercial batch sizes with good productivity (Kis et al., 2020;Skok et al., 2022), largely negating the need for complex intensification.However, there are some developments in this area, mainly limited to academic research (Vetter et al., 2022), but which helps indicate that there may be some desire to intensify these therapies.On the other hand, taking into consideration viral vectors with poor process yields, should a therapy arise that requires a very high therapeutic dose resulting in each batch being able to provide only a few doses, then intensification could be an attractive proposition.Whilst these fields are rapidly evolving, process intensification is not as straightforward compared to protein-based therapeutics.
Lastly, overall analytical technologies for newer modalities may hinder Level 3 process intensification in these areas.Currently, many common analytical assays for viral vectors tend to be slow, low throughput, highly variable with long preparative times (Gimpel et al., 2021;Werle et al., 2021).All of these factors are not compatible with PAT, which requires sensitive assays with very quick turnaround times to enable real-time process changes based on the analytical data.There have been several promising advances recently in F I G U R E 5 Schematic representation of (A) parallel batch multicolumn chromatography (PB-MCC), (B) sequential multicolumn chromatography (S-MCC), and (C) rapid cycling chromatography (RCC).Within PB-MCC, the harvest material is loaded discontinuously across 2 or more columns where no column overloading is occurring.S-MCC involves maximum utilization through continuous column overloading.As breakthrough occurs on one column, a second column begins loading the overflow material from the first column.RCC chromatography maximizes membrane utilization through shortened residence times, allowing for each cycle to be executed in a minimal amount of time in comparison to that of resins.
high-throughput, PAT-compatible assays for viral vectors (Lam et al., 2022;Wang et al., 2019), which could help to support the movement towards Level 3 intensified processing for these products.

| CONCLUSION
Looking at process intensification holistically, whilst there are many different options and levels, it is important that the level of process intensification implemented meets the specific need of the user or process-a single level 1 step may be sufficient, for example, depending on the productivity needs of the process, desired cost savings or ease of implementation.Level 3 intensification and beyond requires significant investment and development time, whereas the transition from Level 0 to Level 1 or 2 allows the user to build on their process sequentially, gradually improving the productivity of their process through islands of intensification.Certain aspects of process intensification can be applied in PD immediately, without the need for large CAPEX expenditure, while other aspects can realize large OPEX or CAPEX savings in manufacturing.Ideally, intensification strategies should be developed first at PD phase to generate data and confidence on the technologies before moving to manufacturing.Trying to find the balance with regard to what level of intensification is needed for a process could be considered a key factor in the process intensification journey.Process intensification is not all or nothing.

Figure 5
Figure5details the phases and differences between those of PB-MCC, S-MCC, and RCC.Within this table, both resin-based column chromatography and membrane-based chromatography are evaluated for the capture step and defined as scenarios B1 and D1 for resin and B2 and D2 for membrane-based approaches.To further define these scenarios, sequential multicolumn and sequential multimembrane chromatography take advantage of the diffusive properties of the medium as the load material is continuously loading.The medium is thereby overloaded while the flow-through is captured on Scenarios of downstream chromatography include classical bind and elute (B&E), parallel batch multi-column chromatography without column overloading and discontinuous loading (PB-MCC), batch rapid cycling chromatography with convective diffusive membrane (RCC), sequential multi-column chromatography with column overloading and continuous loading (S-MCC), sequential multi-membrane chromatography with overloading (S-MMC), flow through chromatography (FT), and Continuous Viral Inactivation (Continuous VI).Scenarios B1 and D1 are representative of resin-based chromatography, whereas B2 and D2 represent membrane-based chromatography.Outputs comparing each of the levels of Process Intensification are shown in Figure 4.