Visions for the Future of Biopharma Manufacturing

Published on: 
Pharmaceutical Technology, Pharmaceutical Technology-05-02-2016, Volume 40, Issue 5
Pages: 22–29

Experts discuss some of the emerging trends in bioprocessing in 2016, including 4D bioprinting, 2D-NMR, and the CAR-T design space.

Although there are typically many initiatives in the works that have the potential to advance biopharmaceutical development, there are few that are as innovative (and currently ready to be actively applied to drug manufacturing in practice) as some of the improvements gaining steam in 2016. In 2015, this publication saw an uptick in the use of single-use systems, continuous manufacturing, novel expression systems, upgrades to perfusion equipment and techniques, and new protein biologicals, such as chimeric antigen receptor T-cell (CAR-T) therapies. This year, Pharmaceutical Technology reviewed current academic literature and spoke to experts to present rapidly progressing approaches in biologics manufacturing and analytics, including new concepts on bioprinting, two-dimensional nuclear magnetic resonance spectroscopy (2D-NMR), and the optimization of CAR-T manufacturing.

3D and 4D bioprinting
Three-dimensional (3D) manufacturing is also known in some circles as additive manufacturing. These technologies are mostly being explored for personalized doses, and could serve to help with the drug delivery of protein-based drugs. Ihalainen et al. says there have been recent advancements in the delivery of proteins, macromolecules, and cells. Biomolecules and living cells are also being printed for cell-culture analytical modeling and therapeutic applications (1, 2).

There are many different physical technologies for depositing material in 3D printing. 3D printing examples include “layered drug-delivery constructs” such as oral film formulations (3), syringe extrusion for tablets (4), laser printing, and sound wave printing technologies, among others. For biomolecules, beyond defining a structure, the benefits of printing include increased control, automation, and reproducibility. Scientists are working on ink development for biomolecules, also known as bioinks. Proper viscosity and surface tension of bioinks are concerns-the potential for shear stress on cells and damage from high curing temperatures are risks, and investigators must make sure the flow of starting materials from an ink nozzle remains consistent. The “large-scale manufacturing of biosurfaces” is also a “critical issue for the application of inkjet printing in biomolecule deposition,” according to Ihalainen et al. (5).

Inkjet printing has been used to create a distinct pattern of cells for neuronal adhesion, a microrray biosensor for the “rapid detection of both protein and bacterial analytes under flow conditions” (5), loading of solutions onto an ELISA plate, and using inkjet printing for the automated deposition of collagen for cell adhesion in predetermined patterns.

Thermal inkjets have been shown to deposit enzymes, deliver proteins successfully to silica supports, and print DNA onto hybridization membranes. Other printing techniques exist, but are contact techniques: flexographic, screen, and gravure printing. These techniques have been used for the printing of active antibodies, the manufacture of disposable biosensors, and the deposition of biomolecules, respectively (5).

A potential drawback to bioprinting is that biomolecules are prone to aggregation after unfolding as a result of shear forces that are too high. “In addition,” Ihalanianen et al. write, the “fast-pressure pulse used to generate compression and eject a drop in piezoelectric printers could cause denaturation” (5). William Whitford, strategic solutions leader, bioprocess, GE Healthcare Life Sciences, notes that aside from hydrodynamic force and pressure drops, chemical interactions with a matrix could also be concerns when printing biomolecules.

4D bioprinting. Some experts have recently described 4D bioprinting, which is essentially a version of 3D bioprinting that confers enhanced biological functionality. There are four distinct types of 4D printing, according to Whitford: shape change (smart biopolymer), size change (patch), pattern change (droplets of cells), and biological change (e.g., phenotype) (2). This last type produces changes that are not structural changes, per se, but self-actuated biological changes in response to some sort of stimulation. These changes occur mostly at the cellular level, and not necessarily conformation-related.

Whitford admits there are three schools of thought on bioprinting. First, there is classic 3D printing. Then, there are those who call only change of morphology after printing, 4D printing. Whitford says, however, that the third definition of the “fourth” dimension of bioprinting signifies any cellular development or function that is engineered into the process and occurs after the printing event. The fourth dimension “could be shape, but it could be differentiation of a cell into a functional cell in situ in the tissue or organoid you are printing,” says Whitford. “This self-actuated development, it doesn’t have to be shape-it could be a biological development.” He gives the example of cells aligning after being printed onto a matrix and polarizing so that one side is different than the other.

Printing can only be considered bioprinting, says Whitford, if cells are actually employed within the process. Even if a biological molecule that is printed will be used in a biological application, it may not necessarily be considered bioprinting. However, 3D printing of, for example, a collagen structure or framework that is immediately infused with cells and then implanted in the body is considered by some to be bioprinting, Whitford asserts. Therefore, the hallmark of bioprinting is an engineered, intentional post-printing development. “Say someone prints some cells in a mass and implants them by a kidney and the cells develop into a pancreatic islet structure that starts secreting insulin, for example. Both the shape change into the islet structure and the development of the cell toward a functional secreting cell are considered by some to comprise the fourth dimension,” Whitford explains. “The investigator didn’t print it that way. He or she printed the preliminary materials, the preliminary cells, and some type of matrix-and maybe even some people are including DNA vectors for growth factors. These preliminary materials then self-activate, in a prescribed, defined manner towards the final tissue or structure that functions the way [the investigator] anticipated.”


A 3D-printed ear. Scientists, led by Anthony Atala from the Wake Forest Institute for Regenerative Medicine, introduced the Integrated Tissue and Organ Printing System (ITOP) in 2016. Using this “cell printer”, the cells that were printed were able to retain viability (6). Through the use of cell-laden nutrient hydrogels coupled with biodegradable polymers for strength, Atala and his team were able to grow a human structure in the shape of an ear. The “sacrificial” scaffolding that held the shape of the ear during cell differentiation was designed to dissolve once the structure was strong enough on its own. A limiting factor of tissue replication is usually the vasculature of tissues once they have been transplanted. In previous mouse models, stem cells have turned into human tissue, and cartilage and blood vessels have formed. In the Atala experiment, the investigators engineered in microchannels to help guide nutrient flow and mimic the function of a vasculature network for the printed cells (6).

Other bioprinting applications. Other companies are trying to print cells to eventually form into tissue-like structures that can be later used for ADME-Tox (absorption, distribution, metabolism, and excretion) toxicity studies, drug development, research purposes, organ function modeling, and for human biological disease models. A company called Organovo has a machine called the NovoGen Bioprinting Platform that is capable of printing human 3D tissue; this company teamed up with cosmetics giant L’Oréal in 2015 to develop tissue-like structures and to bioprint skin with the ultimate goal of better understanding how to test L’Oréal’s cosmetics without relying on animal models (7). Procter & Gamble is in a similar venture to test the toxicity of beauty products, and has teamed up with Singapore’s Agency for Science, Technology & Research (A*STAR) for a bioprinting project (8). Poietis and BASF are also working on their own project to improve BASF’s Mimeskin, the company’s high-resolution, laser-printed skin modeling project (9). These ventures could drastically reduce the time it takes to grow skin tissue-from around four weeks in a traditional culture to just several days.

New analytical techniques: (2D-NMR) for the structural analysis of mAbs
New enhancements in the field of NMR spectroscopy have the potential to advance biopharmaceutical quality control. The new use of a 2D-NMR tool can give crucial information to investigators about the conformation of a drug product and allows them to make conclusions about sample similarity (10). 2D-NMR will assist in determining formulations that stabilize the desired structure of a monoclonal antibody (mAb), and can also assist in determining the flexible and rigid regions of mAbs, notes Kimberly L. Colson, PhD, business development manager at Bruker Biospin. Colson adds that NMR is increasingly being deemed an acceptable material validation approach.

There are misconceptions about the complexity of assessing a structural protein through the use of NMR-specifically that NMR is too complicated and “is not adaptable to the quality controlled environment of manufacturing facilities” (11). New work performed by research scientists at the laboratories of the National Institute of Standards and Technology (NIST) (John Marino, Robert Brinson), FDA (David Keire), and Health Canada (HC) (Yves Aubin, Houman Ghasriani) over the past few years, however, has demonstrated that it is possible to map protein therapeutics, including mAbs, by 2D-NMR (at 13C and 15N natural isotopic abundance) without the need for stable-isotopic labeling (12).

Typically, large proteins have nuclei that overlap in NMR, affecting the resonance of the resulting image. “Conventional wisdom had held that NMR measurements begin to fail for molecules that are larger than 30,000 Daltons without using more sophisticated isotope labeling techniques,” commented the researchers. “The manufacture of isotopically enriched ‘labeled’ proteins with improved NMR properties is expensive, low yielding, and not feasible in a biopharmaceutical setting. Both the intact mAb and constituent domains were generally considered by many to be too large to be measured at natural isotopic abundance (e.g., an intact mAb is 150,000 Daltons and the Fc and Fab fragments are 50,000 Daltons).” As a result of this technical restriction, the researchers said, no experiments using 2D-NMR (to their knowledge) were attempted before their work was published. A key advance that facilitated the use of 2D-NMR for the analysis of larger proteins at natural isotopic abundance was the introduction of cryogenic probe technology, said the researchers, which made it possible to yield spectra with much greater sensitivity-as much as four- to five-fold better than probes at room temperature.

Indeed, says Colson, advances in NMR have made structure analysis easier. “Improvements in CryoProbe technology have greatly enhanced the sensitivity of NMR, while the high reproducibility of today’s NMR spectrometers makes comparison of datasets easy, allowing a NMR spectroscopist to detect very minor changes in structure.” Colson adds, “Pulse sequence development enables the technique to be more quantitative while using 2D-NMR Heteronuclear Single Quantum Correlation (HSQC) approaches. In addition, the development of rapid acquisition techniques has also reduced the experiment time.” She says that reproducibility has been enhanced so much that “it can now be expected that nearly identical spectra can be obtained on different instruments.” 


NMR is unique in that it allows for the assignment of signals at atomic resolution, and the 2D spectral map provides a read-out of the primary, secondary, tertiary, and quaternary structure of a protein, say the investigators-and under defined conditions (e.g., temperature, pH, buffer, presence of excipients, etc.), a spectral output will be unique to a protein. Thus, 2D-NMR can reveal crucial information about the biological activity, quality, and safety of a protein: “Any perturbation of the higher-order structure could have an impact on the efficacy or safety of a monoclonal antibody product, and can in principle be detected by the 2D-NMR map. If there are resonance assignments, these spectral deviations can be mapped onto the structure at atomic level resolution to identify where in the structure of the protein has been altered,” the researchers note.

The NIST, FDA, and HC groups attest that besides NMR, no other mAb characterization techniques that are capable of providing high-resolution analysis data can be practically applied to drugs at natural isotopic abundance. Although they mention X-ray crystallography is a high-resolution choice, they say that this method requires a “protein crystal that can diffract X-rays and conditions that are significantly different from the drug formulated in solution.” They continue, “Most other techniques, such as HPLC [high-performance liquid chromatography], SEC [size-exclusion chromatography], DSC [differential scanning calorimetry], FTIR [Fourier transform infrared spectroscopy], and CD [circular dichroism] spectroscopy, afford high sensitivity but lower resolution. They measure gross structural features and can actually miss small structural differences. There are cases of mutant proteins exhibiting a highly similar analytical signature as the reference drug product by these lower-resolution approaches, but then failing a biological assay that looks at its potency.” Thus, the researchers stress, using NMR to generate high-resolution structural data is the best way to guarantee a protein is properly folded. Colson adds that both 2D and 1D approaches will likely be “used in tandem because the overall time to acquire and interpret the data will be small relative to the information provided, while the time needed for other technologies is substantially more.”

Current uses of 2D-NMR: Biosimilarity and beyond. Sponsors of biosimilar products have already incorporated NMR data in their filing to regulatory agencies, say the scientists, and they expect this trend to continue. “From our conversation with industry, the use of 2D-NMR is becoming more widespread, as it is recognized that the 2D map can provide a complete visualization of primary, secondary, tertiary, and quaternary structure map (higher-order structure),” the researchers state. “As 2D-NMR data were part of the submission package for Zarxio (filgrastim-sndz)-an 18,800 Da protein now approved in the US as a biosimilar to the reference licensed drug Neupogen-such data would be expected to be present in future biosimilar submissions for regulatory approval.” If 2D-NMR is to be used as a tool for the direct comparison of biosimilars, however, experimental conditions for the reference molecule have to match those for the proposed biosimilar (10), the researchers stress. “In addition, with NMR solution conditions and data acquisition parameters controlled between samples being compared, multivariate statistical analysis approaches are becoming common methods to describe the variance between the proteins in these samples. These ‘chemometric’ tools hold promise in the development of metrics for comparability assessment.”

The investigators note that in 2015 and 2016, NIST and other groups-such as FDA, Amgen, Pfizer, and Health Canada-published studies on NMR methods for protein therapeutic analysis. Although they say NMR is already a common tool for small-molecule therapeutics, it can also be used in biomanufacturing to “simultaneously monitor many metabolites in the bioreactor making protein therapeutics.” In conclusion, the researchers say, “The demonstration of measuring detailed spectroscopic fingerprints of molecules as big as mAbs will, if not already, trigger a wider application/adaptation of NMR methods (1D and 2D) to the biomanufacturing and bioprocessing.” To harmonize and establish best practices for 2D-NMR, NIST is organizing a large interlaboratory study with 30 partners from nine different countries.

Optimizing manufacturing for the production of cellular therapies
Although the manufacture of chimeric antigen receptor T cells (CAR-T) is a fairly new undertaking, many companies are already looking at ways in which they can optimize production to yield more efficient therapies. There have been a few improvements in the production of T-cell-based therapeutics, notes Christopher A. Klebanoff, MD, assistant clinical investigator at the Center for Cancer Research at National Cancer Institute (NCI), who specializes in both preclinical and clinical development of adoptive T-cell therapies. Klebanoff says the first upgrade in the process has been the “development of largely closed culture systems that require minimal manipulation from the time of apheresis through the expansion and gene transduction of a patient’s T cells.” The use of serum-free culture media, the ability to “reliably” cryo-preserve modified patient T cells for shipment, and first- and second-generation cell subset isolation strategies-“including antibody-based paramagnetic microbeads and reversible streptamers”-have all contributed to enhancements in T-cell manufacturing techniques, comments Klebanoff.

Cell expansion. T cells transduced with chimeric antigen receptors are among some of the most popular types of cell therapies that are under investigation, and many of the studies in the pipeline seek to ascertain which specific subset of T cells is tied to the best long-term, clinical outcome.

“One of the main limitations to improving the efficacy of ACT [adoptive cellular immunotherapy] is the capacity to expand T cells to therapeutic yield while limiting effector differentiation …  current in vitro methods to expand cells to sufficient numbers and still maintain a minimally differentiated phenotype are hindered by the biological coupling of clonal expansion and effector differentiation,” writes Crompton et al. (13). Explains David Chang, MD, PhD, executive vice-president of R&D and chief medical officer at Kite Pharma, “Naïve T cells are mature lymphocytes that have survived to thymus selection, present in the periphery in a resting status, and have never encountered any antigen. Upon antigen stimulation, they undergo differentiation into effector, memory T cells (central and effector).” Specifically, peripheral blood cells are made up of naïve (TN), stem cell memory (TSCM), central memory (TCM), effector memory (TEM), and terminal effector (TEFF) cells (14). As cells undergo differentiation, they lose certain functional abilities-including proliferative and survival potential-and loss of these functions makes cells less therapeutically valuable.


According to Klebanoff et al., TN, TSCM, and TCM subsets are associated with superior T-cell engraftment, persistence, and antitumor immunity (14, 15) compared with TEM and TEFF. Even though it is generally understood that mature cells have more potent cytotoxicity, and younger cells are initially less lytic than TEM and TEFF subsets-Klebanoff says younger subsets “are capable of far more robust in-vivo expansion and persistence compared with the more differentiated subsets-the net result is that a greater number of highly lytic T cells are ultimately delivered to the tumor.”

In 2013, Cieri at al. concluded that stem memory (TSCM) cells displayed the best anti-tumor response in adoptive T-cell therapies when compared with other types of memory cells (15)-and only naïve T cells gave rise to a significant number of TSCM cells-therefore, many newer research efforts may focus on the homing in of these two subsets of T cells that can persist for years, rather than days. On a 2015 fourth-quarter call, Chang revealed that Kite Pharma, a biopharmaceutical company focused on the manufacture of immunotherapies, is focusing its efforts on culturing “more naïve cells, because all the data indicate naïve cells are the ones that expand better and [are] more potent in terms of antitumor activity” (16). Often, following adoptive transfer, there can be issues with the proliferation and persistence of cells, so companies such as Kite are trying to optimize the production of cells with the most desirable phenotype (or as mentioned, those that are minimally differentiated). Chang said in some cases, maintaining effector progeny in a continuous fashion could be achieved by shortening the cell-expansion phase during cell culture. He also noted that using different cytokines (such as IL-2) during the cell manufacturing process or using inhibitors of certain signaling pathways (such as the pharmacologic inhibition of the serine/threonine kinase Akt [17]) could keep cells from differentiating from an optimal, naïve state and could stave off replicative senescence following adoptive transfer. Klebanoff affirms that generation of minimally differentiated subsets “might be accomplished by using alternative cytokines besides IL-2 during the ex-vivo expansion phase, such as IL-15 and IL-21, or physical isolation of minimally differentiated T cell subsets before expansion and gene modification.” He adds, “products from Miltenyi Biotec and Stage Cell Therapeutics have been very enabling for the isolation of defined T-cell subsets.”

Many other experts also suggest that reprogramming and “dedifferentiation” of senescent effector cells back into naïve or memory T-cells could be a viable option and would be capable of producing cells that maintain functional integrity (17, 18).

As investigators continue to take steps to improve the persistence and immunological memory of tumor-infiltrating lymphocytes and genetically engineered T lymphocytes, there are still some challenges ahead. As Klebanoff notes, “generation of standardized cell products, in terms of phenotype and subset composition, remains difficult because of a lack of highly efficient and inexpensive cell isolation reagents.” He explains that improved cell isolation techniques-those that can tease out defined T-cell subsets-would be most beneficial for the field. Regardless of the introduction of new technologies, says Klebanoff, “a critical limitation for the continued progress in the field of T-cell therapy is the identification of receptors reactive against targets expressed by common solid cancers.”

Protein engineering. Other efforts to make CAR-T cells more specific and to limit their toxicity to normal cells have been examined. Work by Rodgers et al. revealed that CAR-T cells can be engineered to be switchable (an sCAR-T), and can be activated when in the presence of engineered cancer-specific antibody-based molecular switches (19). Because the sCART-T can kill in a more controlled manner, the approach is associated with a lower risk of cytokine release syndrome-and the activity of the sCARs can be optimized by varying the valency and geometry of the switches (20).

Another group of investigators pointed out in Cell that when T cells were coupled with antigen-sensing circuits, they could target two antigens at once (much like a bispecific antibody, but in a T cell). The use of engineered SynNotch receptors-an extracellular antigen-recognition domain of an antibody that drives the expression of a CAR for a second antigen-facilitates “therapeutic discrimination,” according to the authors, while also allowing for “immune recognition of a wider range of tumors” (21).

1. I.M. Hutchings and G.D. Martin, Inkjet Technology for Digital Fabrication (Wiley, West Sussex, United Kingdom, 2012).
2. W. Whitford, "Another Perspective on 4D Bioprinting,", accessed on Feb. 20, 2016.
3. M. Preis, J. Breitkreutz, and N. Sandler, Int. J. Pharm. 49, pp. 578–584 (2015).
4. S.A. Khaled et al., Int. J. Pharm. 461, pp. 105–111 (2014).
5. P. Ihalainen, A. Määttänen, and N. Sandler, Int. J. Pharm. 494, pp. 585–592 (2015).
6. H.-W. Kang et al., Nat. Biotechnol. 34, pp. 312–319 (2016).
7. Organovo, "L'Oreal USA Announces Research Partnership with Organovo to Develop 3-D Bioprinted Skin Tissue," Press Release, May 5, 2015.
8. A. Scott, "P&G Funds Project To 3-D Print Human Skin," C&EN 93 (22), p. 6 (June 1, 2015).
9. BASF, "BASF and Poietis Sign a Research and Development Agreement on 3D Bioprinting Technology for Advanced Skin Care Applications," Press Release,, Accessed on Feb. 20, 2016.
10. H. Ghasriani et al., Nat. Biotechnol. 34 (2), pp. 139­–141 (2016).
11. C. Jones, Y. Aubin, and D.I. Freedburg, “Advances in Protein Characterization” supplement to BioPharm Int. 23 (8), (2010).
12. L.W. Arbogast et al., Pharm. Res. 33 (2), pp. 462–475 (February 2016).
13. J.G. Crompton, M. Sukumar, and N.P. Restifo, Immunol. Rev. 257 (1), pp. 264–276 (January 2014).
14. C.A. Klebanoff, L. Gattinoni, and N.P. Restifo, J. Immunother. 35 (9), pp. 651–660 (2012).
15. N. Cieri et al., Blood 121 (4), pp. 573–584 (2013).
16. R. Hernandez, "Kite Pharma Stresses Importance of Cell-Culturing Techniques,", accessed March 2, 2016.
17. J.G. Crompton et al., Cancer Res. 75 (2), pp. 296–305 (Jan. 15, 2015).
18. L. Gattinoni, C.A. Klebanoff, and N.P. Restifo, Nat. Rev. Can. 12, pp. 671–684 (October 2012).
19. D.T. Rodgers et al., Proc. Natl. Acad. Sci. U.S.A. 113 (4), pp. E459–E468 (Jan. 26, 2016).
20. J.S.Y. Ma et al., Proc. Natl. Acad. Sci. U.S.A. 113  (4), pp. E450­–E458 (Jan. 12, 2016).
21. K.T. Roybal et al., Cell 164, pp. 770–779 (Feb. 11, 2016).

Article DetailsPharmaceutical Technology
Vol. 29, No. 5
Pages: 22–29
Citation: When referring to this article, please cite it as R. Hernandez, "Emerging Trends in Biopharmaceutical Manufacturing: 2016, " Pharmaceutical Technology 29 (5) 2016.