Managing particulates in cell therapy: Guidance for best practice

DOMINIC CLARKE1,2,*, JEAN STANTON3,*, DONALD POWERS4, OHAD KARNIELI1,5,SAGI NAHUM6, EYTAN ABRAHAM1,7, JEAN-SEBASTIEN PARISSE1,8 & STEVE OH1,9

1International Society for Cellular Therapy Process and Product Development Subcommittee, USA, 2Charter Medical,Ltd.,Winston-Salem,North Carolina, USA, 3Johnson & Johnson, Philadelphia, Pennsylvania, USA, 4Janssen SupplyChain, Spring House, Pennsylvania, USA, 5Karnieli Ltd,Tivon, Israel, 6PluristemTherapeutics, Haifa, Israel, 7LonzaWalkersville, Inc.,Walkersville,Maryland, USA, 8Aseptic Technologies, Les Isnes, Belgium, and 9Stem Cell Group,BioprocessingTechnology Institute, Centros, Singapore

AbstractThe intent of this article is to provide guidance and recommendations to cell therapy product sponsors (including devel-opers and manufacturers) and their suppliers in the cell therapy industry regarding particulate source, testing, monitoringand methods for control. This information is intended to help all parties characterize the processes that generate particu-lates, understand product impact and provide recommendations to control particulates generated during manufacturing ofcell therapy products.

KeyWords: cell therapy, disposable, injectable, parenteral, particle, particulate

Background and introduction toparticulate challenge

The cell therapy community is currently engaged inmore than 2500 clinical trials around the world, with15% of those being industry-sponsored clinical trialsand the remainder sponsored by leading academiccenters [1]. In addition to commercially available celltherapy products and those in development, cells havebeen used as a standard of care for decades in themedical practices of hematology and oncology. Stemcell transplantation, for example, continues to be rou-tinely used in, and investigated for, an increasinglydiverse and growing list of malignant and nonmalig-nant diseases. More than one million stem celltransplants have been performed globally to date [2].When stem cell transplantation is used as a form ofcancer treatment, the requirements for particulatecontrol in the material used for the transplant havebeen limited to aseptic techniques used during pro-duction, visual inspections at release and immediatelybefore administration and in-line filters utilized at thebedside.

Cell therapy products do not undergo final filtra-tion steps (e.g., sterile filtration) as is routinely donefor many biologic products; consequently, compa-nies developing commercial cell therapy products and

suppliers need to consider the potential risk of par-ticulate matter contamination from the externalenvironment, during the manufacturing process orrelated to cellular or protein degradations.The pres-ence of particulates in cell therapy products currentlyrepresents a quality challenge and may pose safety con-cerns.The challenges of particulates include difficultiesidentifying the nature of the particulates found in thefinal or injected product, limited or no understand-ing of the impact of particulates on the cells anddifficulty defining relative contributions of differenttypes of particulates to any in vivo toxicity or immu-nogenicity effects. Together these issues mean thatcontrol of particulates is likely to require a combina-tion of analytical methods [3].

The current pipeline of cell-based therapies rep-resents a maturation of the science surrounding theseproducts. As such, the next generation of productscarries enhanced expectations of quality, safety, effi-cacy and commercial viability. Compared with the firstgeneration of cell therapies, the products currently inclinical development must be better characterized tounderstand the contribution and potential impact ofparticulates.

A number of articles addressing particulate matterexist for the pharmaceutical and bioprocessing indus-tries [4–6], but references to particulates and cell

*These authors contributed equally to this work.Correspondence:Dominic Clarke, PhD, Charter Medical, Ltd., 3948-AWestpoint Blvd.,Winston-Salem, NC 27284, USA. E-mail: dclarke@chartermedical.com

ARTICLE IN PRESS

(Received 5 April 2016; accepted 13 May 2016)

ISSN 1465-3249 Copyright © 2016 International Society for Cellular Therapy. Published by Elsevier Inc. All rights reserved.http://dx.doi.org/10.1016/j.jcyt.2016.05.011

Cytotherapy, 2016; ■■: ■■–■■

therapies have been limited [7]. As highlighted byClarke et al., there are many reasons the regulations,guidelines and practices used to manufacture tradi-tional pharmaceutical drugs or biologic products donot apply equally well for cell therapy products.Tra-ditional injectable products are sterile filtered beforethe final fill.These clarified solutions are much easierto assess with current visual inspection methods [7].Efforts have also more recently been held directly orindirectly by academia with research and development–driven competences to aid in addressing the particulatecontamination issue for cell therapies. One example,is the implementation of bioprocessing integrated strat-egies as mentioned by Cunha et al., in which processintegration can reduce the potential source of per-sonnel contamination [8].

To build on the general awareness brought forthin Clarke’s publication, this article describes the currentchallenges and understanding in the industry. Chal-lenges discussed here include risks to patients, risksto product quality and compliance or regulatory risks.Current understanding includes particulate charac-teristics, measurement methods, composition andsources of particulates that can be discovered in thefinal formulation of a cell therapy product.This articlealso provides guidance and recommendations for el-ements to be included in particulate managementprograms for suppliers and companies developing celltherapies.The information is intended to assist com-panies with their characterization of the manufacturingprocesses and products to minimize particulate con-tamination in the final products and suggests a meansfor suppliers and companies developing cell therapyproducts to work toward improving manufacturingquality and minimizing risks to products and patients.

Particulate risks to patients and products

It is critical to the success of cell therapy products indevelopment that they be both safe and efficaciouswhen used to treat a specific indication in relevantpatient populations. Goals of product development areto understand the risks and benefits specific to the in-tended use of the product and, ultimately, to reducethe risks as the product evolves. Particulate risks relatedto a product’s final formulation are important for manyreasons. The most obvious risk is the potential foradverse events due to particulates that occur after aproduct is administered to a patient.The second riskis the impact of particulates on the product qualityitself, both during production and in a product’s finalformulation. Particulates discovered in final prod-ucts also increase a product’s risk of being recalled,leading to potential clinical trial delays or failure tomaintain commercial inventory.

Patient safety and medical risks

Although the expectation is always to produce safe andeffective products, a variety of criteria should be evalu-ated when trying to determine possible patient risks.Despite the different sources and composition of par-ticulates, there are several common types of pathogenicmechanisms for potential harm to patients. Thesemechanisms include inflammation due to infectionscaused by viable organisms, inflammatory responsescaused directly or through associate leachates thattrigger direct tissue injury, normal and abnormalimmune responses to cellular debris, and tissue damagefrom thromboembolism [9]. Most of the adverse eventsin the literature related to particulates in parenteraldrugs are based on animal studies, in vitro studies orhuman case reports. The nature of the injury anddegree of risk depends on several factors, includingthe route of administration, frequency, particle size andnumber, particle composition and patient popula-tion [4].

The routes of administration mainly considered forthis discussion are intravenous (IV), intra-articular andintrathecal because most cell therapy products are ad-ministered via these routes. The size of the veinschanges as blood flows to the heart and then to thelungs. Most particulates administered intravenouslywill follow this route until they are trapped in the pul-monary capillaries within the lungs, the diameter ofwhich is approximately 12–15 μm. Many of the cellsused for cell therapy purposes are often greater than20–30 μm in size. Their size often leads to signifi-cant entrapment. Unlike many of the intrinsic andextrinsic noncellular particulates discussed in thisarticle, many of the cells will eventually pass throughafter multiple passes by the blood through the lungs.Cells are more flexible when it comes to movingthrough smaller diameters through amechanism called“deformation.” Inhibitors of cellular adhesion mol-ecules (CD49d) on the cell have also been created toreduce the number of passes it takes for cells to clearthe lungs [10].The most common consequences fromthese trapped intrinsic and extrinsic noncellular par-ticulates are compromised oxygen transfer and impairedrespiratory function. Another clinical complication po-tentially resulting from IV or intra-arterial productadministration is granuloma formation, with symp-toms of dyspnea and reduced pulmonary function. Forexample, a report by Garvan describes postmortempulmonary vasculature granulomas due to cellulosefibers. In another report, a large pulmonary granu-loma formed in a patient when a particle becamelodged in an arteriolar wall and eventually erodedthrough to form a giant cell granuloma [9]. Granu-lomas in organs other than the lungs are found to beof negligible safety risk for patients [11]. Another

ARTICLE IN PRESS

2 D. Clarke et al.

potential risk is related to intra-arterial administra-tion of small particulates as opposed to largerparticulates. Both particulate sizes can cause occlu-sion in the vessels, but smaller particulates are moredetrimental to patients as they are capable of causingarteriole occlusions, which can lead to limb-threateningischemia [12].

Case reports related to large particulates are notcommon, most likely because these particulates areidentified during visual inspection and the units arerejected from the lot. However, large particulates maycause direct traumatic damage to the vein itself, as wellas infection if the particulate is not sterile. Small par-ticulates are generally phagocytized by macrophagesand taken up by several organs of the body. Normal-ly, if the amount of particulate is small, then there areno clinical consequences because most organs havereserves to manage the deposits. When the particu-late load is high, however, it may affect patient health,often overloading the spleen and liver and causing in-fections [13]. Several studies have also suggested thatthe shape of the particulate matters. One such studyshowed that rabbits injected with cellulose died within4 min of administration due to acute toxin response.In contrast, rabbits injected with cellulose spheres didnot suffer the same result [7].

The number of particulates administered has beenconsidered indirectly through clinical studies thatlooked at adult patients in the intensive care unit.Thesestudies were conducted to evaluate the use of IV in-line filters.The results indicate that use of these filtersreduced the incidence and time of onset for particulate-induced phlebitis. An inference from these studies isthat the level of risk increases as the particulate loadincreases [14].

Particulate composition is another factor with a po-tential to cause harm. Barber provides an overview ofseveral animal studies that tested different particu-late compositions, including glass, fiber and polystyrene[15].The clinical effects ranged from minor to serious(local inflammation, granuloma and death).The par-ticulates responsible for the more serious reactions wereplastics, fiber paper and polystyrene. Patient-relatedsequelae attributed directly to glass particulates includepulmonary granulomas, phlebitis, adult respiratory dis-tress syndrome and systemic inflammatory response[16].

Metal particulates are seen as one of the most dan-gerous contaminants identified in the literature, bothfor IV therapies and nutritionals. Stainless steel is mostcommonly associated with the pharmaceutical indus-try, whereas metal particulates will likely be observedmuch less frequently in cell therapies. Literaturesearches found no animal or in vitro studies associ-ated with stainless steel. The records found wereassociated with product recalls [17]. Inherent par-

ticulates such as cellular debris or proteinasecomponents have the risk of triggering an unin-tended immune response, as can be seen in theliterature for both the blood industry and therapeu-tic proteins. Cell therapy products that are stored forany period of time outside of cryogenic tempera-tures run the risk of containing high levels of cytokines.Cytokines have the potential of activating pro-inflammatory pathways, as well as causing extravasculartypes of reactions [18]. Proteinaceous particulate matteralso poses a risk of triggering host responses, result-ing in antibodies to proteins expressed on cellularsurfaces.

The patient populations most at risk for particulate-related sequelae include patients with existing tissuedamage, critically ill patients and neonates. Becausethe majority of cell therapy products in developmentare for adult indications, neonatal populations will gen-erally be excluded from this discussion. Cell therapyproducts are indicated for patients with specific tissuedamage, such as spinal cord injuries and cardiacdisease. In both cases, the products are used to rescue,replace and help repair the site of injury or infarct.Particulates passively administered to these patientswith cell therapy products could potentially have majorsecondary sequelae at the damaged site.

Another critical patient population to consider whenit comes to particulates is oncology patients. Many ofthese patients have impaired immune systems, eitheras a direct result of the disease or secondarily throughtreatments. If the particulates are nonsterile, then pa-tients are at potential risk for infections. Particulatesalso have the potential to have a negative impact ona patient’s immune system through possible immunesuppression or by activating the immune systemabnormally.

Evaluating each cell therapy product requires a fullunderstanding and identification of various factors; onesize does not fit all. Risk assessment is commonly per-formed in accordance with recognized guidancedocuments and standards, with risk generally beingdefined as the combination of the probability of oc-currence of harm and the severity of that harm. Table Iis a visible depiction of some of the criteria de-scribed previously, which can be used to assess potentialrisks to the patient. Given the unique nature of manycell therapy products, especially autologous prod-ucts, the risk of limited product availability should alsobe considered when assessing overall risk and partic-ulate impact.

Product risk

Cell therapy products face a unique challenge as thecells themselves are the final product. Thus, there isan inherent limitation to the purification processes that

ARTICLE IN PRESS

Managing particulates in cell therapy 3

can be performed, and effective removal of particu-lates from the final formulations will be limited. It isthe responsibility of the sponsor of the cell therapyproduct to characterize and understand the load,source, composition and potential impact of particu-lates in the final formulation of their product.

The impact of inert particulates on cells and cellcultures will vary greatly depending on the proper-ties of the particulate and of the cell lines. Cellularadhesion can be affected by exposure to particu-lates, depending on whether the particulate is takenup by the cell and the basic topography of the par-ticulate.When taken up by cells, specific particulatesmay affect a cell’s cytoskeleton, potentially affectingcellular adhesion. Several studies with silicon-basedparticles have demonstrated an induced membrane de-formation when the silicon-based particulates comein contact with human-derived cells, leading to cel-lular lysis [19]. Inert particulates that remain externalto the cellular membrane can have opposite effects oncellular adhesion. First, particulates can provide asurface for cellular adhesion and thus promote ag-gregate formation. Cellular aggregates can clog deliverysystems and negatively affect the mechanism of actionand efficacy of a product. A particulate’s surface fea-tures can also reduce cellular adhesion. Studies haveshown that rougher surfaces are accompanied bychanges in cell physiology such as integrin expres-sion.This altered cell architecture has been shown toresult in decreased synthesis of extracellular matrix(ECM) proteins, which are crucial to cellular adhe-sion [20].

Upon exposure to cells, particulates can be en-gulfed by cells through endocytosis. The act ofendocytosis can have an impact on a cell and its en-vironment in many ways. In addition to affectingcellular adhesion, endocytosis can also be an energydrain on the cell, increasing both respiration and met-abolic activity [21]. In one study, metallic particulateswere shown to be cytotoxic [22]. Another study dem-onstrated that presentation of processed particulateantigens by antigen presenting cells can lead to theinadvertent activation of cells via release of media-

tors into their environment [23]. Moreover, particulatestaken up by cells can be released into the cytoplasm,which can result in cytotoxic effects due to energydepletion, membrane disruption and organelledestruction.

Particulates can also be detrimental to the viabil-ity and functionality of the cell culture. Leachables andextractables from these types of particles can alter thepH of the environment or produce compounds thatare toxic to the cells—both immediately and overtime—which could directly affect product stability.

Under the heading of product risks, it is impor-tant to include the risk of product recall.The presenceof foreign visible or sub-visible particulate matter inparenteral formulations has been one of the mostcommon reasons for product recalls. During the period2008–2012, the U.S. Food and Drug Administra-tion (FDA) reported that 22% of recalls for sterileinjectable drugs were caused by the presence of visibleparticles [17]. As discussed earlier in this section, thepresence of certain types or numbers of particulatematter can pose health hazards. Health authoritiesrequire that they be notified of product recalls andadverse events related to products, either clinically orcommercially. Therefore, it is prudent for manufac-turers and their suppliers to work diligently towardcharacterizing visible and sub-visible particulates fortheir products. The amount of unexpected particu-late that is present in parenteral products is anindication of the pharmaceutical quality of the product.Some of the themes the FDA noted in the 483s issuedto manufacturers in 2012 related to particulate matterand to the establishment of a maximum allowable rejectrate, training and certification of personnel perform-ing visual inspections; the need to conduct a thoroughinvestigation regarding particulate matter; and iden-tification of sources during production.

Particulate characterization

The type, size, quantity, composition and source ofparticulates are descriptions used in the pharmaceu-tical industry to characterize process and product

Table I. Cell therapy particulate risk to patient.

Risk level

Category Low Moderate High

Patient type Adult InfantInfusion route Subcutaneous Intraocular Intravenous

Intramuscular Intrathecal Intra-articularInfusion frequency Once ManyParticulate type Inherent Intrinsic ExtrinsicParticulate size <10 μm <100 μm >100 μmNo. of particulates None Many

ARTICLE IN PRESS

4 D. Clarke et al.

particulates. Particulates are classified as inherent, in-trinsic or extrinsic as described in Table II [9]. Inherentparticulates are defined as materials that are ex-pected from the product formulation and are typicallyan accepted product characteristic. Intrinsic particu-lates are defined as materials that arise from sourcesrelated to the product formulation, packaging and pro-cessing. Extrinsic particulates are defined as materialsthat do not originate from product formulation, pack-aging or processing but are of foreign or unexpectedsources [6].

As introduced earlier, particulate size is common-ly categorized as being visible or sub-visible.The generalconsensus within various existing pharmacopeia mono-graphs has established visible particulates as thosegreater than 100 μm. With respect to the acceptedquantity of particulates, the U.S. Pharmacopeia USP<1 > Injections “essentially free” along with the Euro-pean “practically free” and Japanese “free from readilydetectable” Pharmacopeias (EP 2.9.20 and JP 6.06,respectively) [24–26] have similar descriptionsand requirements for visible particulates. USP<788 > Particulate Matter in Injections is generally es-tablished as a means for setting limits on the size andnumber of particulates based on container volumes,and these limits are harmonized with EP and JP [27].Currently, suppliers and cell therapy sponsors use thesedocuments as guidance, but critical challenges exist.

Particulate evaluation methods

It should be noted that this section is not intended toaddress and describe in detail all particulate detec-tion or characterization methods; rather, it provides asummary of some of these methods. A variety ofmethods and technologies are available to measure andidentify particulates from the visible to the sub-visible

level, and they have been previously described at lengthfor many applications [5,7,9]. Furthermore, techno-logical advances have been relatively limited, so thisrepresents an area for future development.

Before any mitigation or particulate reductionsteps can be assessed and implemented for a specificcell therapy product, process and product character-ization is a priority. Particulate evaluation andmeasurement is commonly performed as part of thecell therapy manufacturing release process. Althougha number of techniques exist, visual inspection is themethod of inspection most commonly used by bothsuppliers and sponsors.

Different factors and criteria need to be consid-ered to determine which particulate methods to use.The type of product and its intended application willinfluence the particulate requirements and methods.In general, methods can be separated into either quan-tification or characterization. Table III depicts someof the more common methods, but it does not rep-resent all possible methods available. Some type ofquantification is almost always performed as part ofthe qualification and release criteria of cell therapy pro-duction, both from the supplier and the cell therapysponsor. Quantifying particulates is performed moreoften, given that some of the methods available arenondestructive. Visual inspection allows for the de-tection and quantification of larger particulates. Fordetection of smaller particulates, other methods arerequired such as USP <788 > and ASTM F24-09 [28].These methods are destructive in nature and cannotbe applied to each final product; they are often per-formed along with the particulate characterizationtesting as part of the process and product validation.Aside from quantifying particulates, characterizationof particulates is important to understanding the type,source and composition of the particulates.

Table II. Types of particulates.

Type Definition

Inherent Expected from the drug formulation and thus represent a generally accepted characteristic of the product, within limitsExamples: cell clumps, excipients, etc.

Intrinsic Related to the formulation, packaging or assembly processesExamples: glass, plastic, rubber, etc.

Extrinsic Foreign and unexpected, not part of the formulation, package or assembly processExamples: hair, fibers, paint, etc.

Table III. Common particulate evaluation methods.

Quantification Characterization (all destructive)

Visible inspection (nondestructive) Scanning electron microscopy (SEM)TAPPI (nondestructive) Energy dispersive x-ray spectrometry (EDX)USP <788 > (destructive) Magnetic resonance spectroscopyASTM F24-09 (destructive) Micro-Fourier transform infrared spectroscopy (FTIR)

ARTICLE IN PRESS

Managing particulates in cell therapy 5

Particulate composition and source

Recognizing the source of particulates in combina-tion with establishing particulate composition canultimately aid in identifying potential risks to both thecell therapy product and the intended recipient. Speak-ing in general terms, the sources of contaminantparticles found in clean rooms used in the manufac-turing of pharmaceutical products can be sorted intothe same categories found in other industries [15]:

• Air handling and filtration system• Particles shed by the room construction materials• Personnel activity• Production materials (including packaging)• Equipment and instrumentation

On the basis of the literature, these same catego-ries can be extended to production of cell therapyproducts. Each of the preceding sources contributesproportionally to the overall particulate load of the finalproduct and to the level of risk. A number of authorshave identified proportional contributions in cleanareas. Figure 1 was created using the proportions pro-vided by Barber [15]. Extrinsic particulates, specificallythose from unknown sources, represent the greatestconcern. Additionally, the challenge for product de-velopers is that cell therapy products contain cells, andthe majority of cell therapy products are not “clear”solutions, which makes commonly applied particu-late inspections difficult to perform. This challengepoints to the need to understand and characterizeparticulates for each cell therapy product and manu-facturing process, which should aid in assessing risk

and help to identify locations and strategies for par-ticulate controls.

Particulate matter in cell therapy products can begenerated from a variety of sources within the manu-facturing and production processes, as is the case withboth biological and drug manufacturing.The genericcell therapy manufacturing process flow diagram de-picted in Figure 2 illustrates such a process, includingmultiple handling and processing steps.This figure re-inforces earlier statements and highlights the largenumber of labor-intensive steps that commonly occurin this type of manufacturing. Personnel and their ac-tivities are generally the most significant sources ofnonviable particles in controlled environments. On thebasis of the rationale that people contribute highnumbers of particulates and interact with materials,equipment and other sources, the highest particu-late levels in the clean room will occur when the processis operating with people present [29].

As seen in the Austin Contamination Index(Table IV) from Dr. Philip Austin’s Encyclopedia ofCleanrooms,Bio-Clean Rooms and Aseptic Areas, humansemit huge amounts of potentially hazardous par-ticles even while sitting and not moving [30]. Giventhe nature and complexity of cell therapy processes,particulates have potential ingress routes at every step.Health authorities expect that manufacturers are knowl-edgeable about the particulates and their sources, aswell as their potential ingress routes.To further com-plicate matters, only a few products have actuallyreached commercialization, limiting the collective ex-perience and data that can be disseminated and usedby the industry to better understand their processesand improve the quality of their products.

Figure 1. Contributing sources of contaminant particulates found in clean rooms used in the manufacturing of pharmaceutical products.

ARTICLE IN PRESS

6 D. Clarke et al.

Other common sources of particulates are the ma-terials used in production. For the purposes of thisarticle, materials include the starting materials or cells,raw materials, excipients, disposables and primarycontact materials required for manufacturing. Figure 3illustrates some of the more common sources of par-ticulates both from the material suppliers and the celltherapy sponsor [5]. Given the nature of cell therapy

products and the manufacturing process, particu-lates are usually additive and can accumulatethroughout the manufacturing process, as is demon-strated in Figure 3. Particulates accumulate on thesupplier side of the process with each step pre-formed, resulting in the final container or assembly.If these particulates are not controlled and removed,the particulate load is then transferred to the cell

Figure 2. Generic cell therapy manufacturing process flow diagram illustrating the multiple handling and processing steps.

Table IV. Austin Contamination Index: particles ≥0.3 μm emitted/min in garment indicated.

Personnel activitySnapsmock

Standardcoverall

2-Piececoverall

Tyvekcoverall

Membranecoverall

No movement 100 000 10 000 4 000 1 000 10Light movement 500 000 50 000 20 000 5 000 50Heavy movement 1 000 000 100 000 40 000 10 000 100Change position 2 500 000 250 000 100 000 25 000 250Slow walk 5 000 000 500 000 200 000 50 000 500

ARTICLE IN PRESS

Managing particulates in cell therapy 7

therapy sponsor, where additional accumulation occursas part of the manufacturing process (as indicated bythe arrows in Figure 3).

One way to visualize the process inputs with respectto particulates is through the use of a quality tool, suchas the fishbone diagram (Ishikawa diagram). Figure 4illustrates how the fishbone diagram can be used toevaluate contributing factors or sources of particu-lates.The contributing factors can be evaluated fromthe supplier’s and sponsor’s perspectives. Contribu-tions can be broken down into two categories: suppliersand the cell therapy manufacturing process.These cat-egories can be further divided into additional sourcesof potential contribution, with each being affected byfive main sources:

• Machine• Methods• Measurements• Environment• People

For example, considering the generic cell therapymanufacturing process depicted in Figure 2, a singlestep such as the wash/buffer exchange can be high-lighted and analyzed for potential particulatecontribution using the fishbone diagram. Breakingdown the manufacturing step into the main contrib-uting sources, probable sources can be mapped out

as demonstrated in TableV. Particulate types can thenbe categorized as inherent, intrinsic or extrinsic. Furthercharacterization can include the identification ofpotential ingress routes for particulates. One way to

Figure 3. Common sources of particulates both from the material suppliers and the cell therapy sponsor which can result in an additiveaccumulation.

Figure 4. Fishbone diagram used to depict contributing factors orsources of particulates from the supplier’s (A) and sponsor’s (B)perspectives.

ARTICLE IN PRESS

8 D. Clarke et al.

understand possible failure modes or potential risksassociated with the particulates is to conduct a failuremodes and effects analysis (FMEA), which can providea systematic inductive risks analysis of a system.Thismethod can assist suppliers and sponsors in priori-tizing strategic planning with regards to any studiesor experiments necessary to further define sources andpotential ingress routes of particulates and, from there,to determine potential control measures [31].

Table VI provides an example of such an analysisfor the same step from Figure 2, used to demon-strate the fishbone diagram for the wash/bufferexchange step. As the table indicates, once a suppli-er or cell therapy sponsor has identified the failuresand current controls, they can make decisions regard-ing any additional controls and/or mitigations needed(see Appendix, which describes a case study for par-ticulate mitigation of a cell therapy product). It isapparent from the information described thus far thatparticulates can be found in most cell therapy prod-ucts and that methods exist to help manage thepotential risk.

Particulate guidance and mitigation solutions

Cell therapy products will almost certainly have somelevel of particulate contamination due to the sourceof the starting material and the complexity of theprocess. The cell therapy sponsor has the responsi-bility of understanding the particulate profile of thefinal product (number, size, composition and source)and assuring its safety for the intended patient. Ul-timately, it is the responsibility of both suppliers andsponsors to control and minimize the introduction ofparticulates. Although removal of all particulates is un-likely, each step can be assessed for particulate controlby evaluating the sources described in the prior figures.

All parties are involved in the control and reduc-tion or elimination of particulates, and alignment isimperative. Suppliers are responsible for knowing whereparticulates originate and must also be aware of theend user’s needs; this may be other suppliers or celltherapy sponsors. Depending on the life cycle of a spe-

cific component, the particulate requirements andcontrols may be vastly different. Medical-grade com-ponents will typically have significantly greater controlsin place for particulate management when com-pared with research-grade materials. Although manycomponents are manufactured in clean roomenvironments—either an ISO 8/Grade D/Class 100,000or an ISO 7/Grade C/Class 10,000—some compo-nents or stages of the component’s life cycle mayactually be manufactured outside of a clean room [5].

As part of cell therapy manufacturing, the occur-rence of some particulate matter is currently challengingto avoid. Health authorities expect the product’ssponsor to understand the amount, composition andsource of particulate matter in the final product andthat the knowledge on the topic increase over timeduring development.The end point of this evolutionis that the commercial product will be well charac-terized for particulate matter. Although particulatesare assessed on a regular basis, particulate accep-tance standards specific to the cell therapy industrydo not currently exist. Accordingly, more guidance inthis burgeoning industry could prove useful in the fol-lowing areas: particulate matter characterization, controlof particulate matter contamination, and confirma-tion or verification of a control’s acceptable qualitylimit.

Characterizing particulate load

As described in an earlier section of this article, par-ticulate characterization includes the sources of andingress routes for particulates (process) and the com-position, morphology, size and amount (product).Several quality tools are available to identify process-and product-related particulate matter; these includeprocess flow, fishbone diagram and FMEA.The processflow tool (Figure 2) is a picture of the separate stepsof a process in sequential order. A process flow pro-vides a visual description of the details included in aprocess, including potential sources of particulates gen-erated during the production of a product, as well asthe means to identify, detect and measure particles [31].

Table V. Possible particulate contributions from common wash/buffer exchange.

Contribution Machine Methods Material People Environment

Source PumpWelderCentrifuge

PumpingWeldingCentrifugeFilterSpiking

CellsReagent/mediumContainersTubingFilters

GownsMasksGlovesSkinHair

Clean roomSafety cabinet

Type MetallicRubber

PlasticSynthetic fiberMetallic

DNA; ECM mineralsProtein aggregatePlastic fragmentsFibers; glass

Organic fiberCelluloseSynthetic fiber

Organic fiberCardboard

ARTICLE IN PRESS

Managing particulates in cell therapy 9

The exercise of creating process flow diagrams (PFDs),which are critical in the generation of standard op-erating procedures and batch records, may simply bea brainstorming session. The information generatedfrom the exercise can be used to populate the nexttool, which is a fishbone diagram. This tool is com-monly used to prevent defects or to identify potentialfactors that cause a defect. Taking each of the indi-vidual steps and its specific elements from the processflow exercise, the fishbone diagram can be used tofurther break down the process to provide added insightinto potential sources of particulates. The causes ofa defect can then be grouped into categories, re-ferred to as the five M’s, as depicted in TableVI. Fromthe fishbone diagram, the information can then be usedto initiate a FMEA. Upon completion of a FMEA, asponsor can describe both process and product char-acterization of particulate matter (see Appendix, whichdescribes a process for particulate mitigation of a celltherapy product).

Manufacturing controls

Controlling particulates is as important as address-ing methods to reduce particulates for cell therapies.Vigilance is necessary from the choice of raw mate-rials to the delivery of the final product.The first stepin control and reduction comes from the alignmentof suppliers and sponsors, with understanding of re-quirements versus limitations. Numerous steps can beimplemented to reduce particulate load throughoutthe life cycle, and the main factors involve the par-ticulate sources described previously. Starting with thesources of particulates identified in Figure 1, con-trols can be described as mechanisms that act asbarriers to particulates, being able to deny access tothe products or control particles on components usedin production. Examples of barriers are provided inTable VII.

Control measures needed to prevent particulatecontamination in a product can be separated intotwo groups: current or existing controls versus thosedeveloped as part of efforts to remediate risksidentified in a FMEA. Controls included as part ofroutine current Good Manufacturing Practice in anaseptic manufacturing environment include thefollowing:

• Classification of the manufacturing area• Positive pressure environments (airflow from clean

to less clean areas)• Airflow patterns and airflow velocity• Filtered air (high-efficiency particulate arrestance

[HEPA] or other)• Suitable facilities, areas, equipment, and materials• Trained personnelT

able

VI.

Exa

mpl

eri

skan

alys

is(F

ME

A)

ofw

ash/

buff

erex

chan

gest

ep.

Pro

cess

step

Pot

enti

alfa

ilure

mod

e

Pot

enti

alef

fect

(s)

offa

ilure

Sev

erit

y

Pot

enti

alC

ause

(s)

offa

ilure

Occ

urre

nce

Cur

rent

cont

rols

Det

ecti

onR

PN

Rec

omm

ende

dac

tion

sA

ctio

nta

ken

Sev

erit

yO

ccur

renc

eD

etec

tion

RP

N

Was

han

dbu

ffer

exch

ange

Par

ticl

esin

buff

erm

atri

x

Vis

ible

and

subv

isib

lepa

rtic

les

adde

dto

the

proc

ess

via

was

hst

eps

9S

uppl

ier

man

ufac

turi

ngpr

oces

s

5S

uppl

ier

uses

0.2-μm

filtr

atio

n5

225

Add

addi

tion

alfil

trat

ion

step

sw

hen

buff

erad

ded

topr

oces

s

Filt

rati

onst

epad

ded

91

545

Tub

ing

Tub

ing

cont

ains

part

icle

s9

Tub

ing

shed

ding

duri

ngpu

mpi

ng

4U

selo

w-s

hedd

ing

tubi

ng3

108

Add

rins

ing

proc

edur

efo

rtu

bing

Pro

cedu

recr

eate

d9

23

54

Env

iron

men

tP

arti

cles

intr

oduc

edto

proc

ess

from

envi

ronm

ent

9O

pen

proc

ess

step

s1

Ope

npr

oces

sing

cond

ucte

din

Gra

deA

bios

afet

yca

bine

t

17

NA

NA

71

17

Rou

tine

and

in-p

roce

ssen

viro

nmen

tal

mon

itor

ing

NA

,no

tap

plic

able

;R

PN

,ri

skpr

iori

tynu

mbe

r.

ARTICLE IN PRESS

10 D. Clarke et al.

• Adequate transport and storage• Production flows (product, personnel, equip-

ment, materials, waste, quality control)• Cleaning

Verifying effectiveness of controls

Confirmation or verification of a control’s effective-ness enables manufacturers to assess the value of theprocess, procedures and equipment, as well as to directfurther efforts in this regard. These verification ac-tivities include the following:

• Environmental monitoring• Visual inspection process• Filter testing (process, HEPA, etc.)• Water/placebo runs• Product quality complaints

Environmental monitoring programs verify mul-tiple systems that can affect particulate matter controlin a manufacturing facility, including cleaning/sanitization, personnel and facility/utilities. It is the lastsystem regarding levels of nonviable contaminants orthe particle burden in the air of a process room or pro-duction facility that verifies particle controls. Viableand nonviable particles are usually counted and sizedwith the use of light-scattering particle-counting in-struments, also called optical particle counters. Theobjectives of such a program should include estab-lishing suitable test limits and frequencies andconducting routine monitoring of air quality to de-termine the need for maintenance or filter testing andperiodic evaluations to assess the overall effective-ness in maintaining suitable conditions.

Visual inspection has long been considered prob-lematic in the biologics industry due to the heavydependence on trained personnel to perform the anal-ysis, leading to much subjectivity. For cell therapyproducts, the process becomes even more problem-atic, as was stated earlier in this article. For sponsorsof cell therapy products, it will be prudent to developa visual inspection procedure that can detect visibleparticulates within an opaque liquid.This process orprocedure should include decision making for accept-ing or rejecting an individual product (autologous cells)

or an entire batch (allogeneic cells), depending on thetype of cell therapy product.

There are multiple types of filters used in a par-ticle management program, one of which is a processfilter used to remove nonviable and viable particles fromliquids added to the process during production. Processfilters should be tested after use to ensure they re-mained integral through their use. For cell therapy,these large porosity product filters (e.g., greater than50 μm) can be used for removal of large particulatesor aggregates. Additional filter types include air filters,which lower particle burden in a manufacturing areathrough the removal of particles from the air. Attain-ing cleanliness in a clean room requires air filtrationat a high level of efficiency.There are several types ofair filters, the most well known being the HEPA.Theseair filters must be tested periodically to check forleakage due to damage to the medium or air bypass-ing the filter seal.

After the mitigation plans have been executed andremediation efforts are in place, sponsors shouldconduct a placebo or water run.These runs are con-ducted to evaluate the whole process from end to end,and they use worst-case process conditions, includ-ing those that present the greatest potential for particlegeneration. It is recommended to run the process underthe same manufacturing conditions that are plannedfor the product. Some common parameters to con-sider include the following [6]:

• Mixing speeds• Number of connections made during production/

exposure to the environment• Clamping or valve use• Pumping• Rinsing, flushing and wash steps• General handling practices

As part of a material controls program, sponsorsof cell therapy products should include a process toqualify both materials and suppliers. Material quali-fication includes determining its fitness for use inproduction and how it was produced. As stated earlier,the grade of the material provides some indication re-garding the controls in place during its manufacturing,with quality increasing from research grade to medical

Table VII. Barriers to particles by source.

Personnel Equipment Workstation/environment Materials Air Process flow

Gowning Qualification Access control Qualification Filtration UnidirectionalTraining Cleaning Physical enclosures Filtration Directional airflow Facility designQualification Area classifications Handling Pressure differential Procedural boundariesMonitor Cleaning

ARTICLE IN PRESS

Managing particulates in cell therapy 11

grade.To verify a material’s quality, suppliers also needto be evaluated by a sponsor. Qualification of suppli-ers should include an onsite audit, if possible, whenthe sponsor has no history with the supplier. Suppli-er audits and qualification should be performed as earlyas possible to avoid any unknown challenges that mightinfluence availability and commercialization.

Supplier particulate verification, controland mitigation

Verification, control, and mitigation of particulates forcell therapy products require the combined efforts ofsuppliers and manufacturers, and a number of pos-sible solutions were previously described [6]. From asupplier standpoint, general methods can beemployed—through FMEA, for example—to identi-fy potential ingress routes and to establish controls tominimize particulates. Many of the same methods de-scribed in the earlier sections on particulatecharacterization and manufacturing controls apply.These analysis methods will again point to the con-tribution sources illustrated in the previous section(Figure 4A), and from there one can identify next steps.

For example, consider a general single-use storagecontainer and tubing assembly and evaluate themethods that can be employed by a supplier to verify,control, and mitigate potential particulates, bearing inmind the cleanliness of the manufacturing. Using themethods described in the previous section, manufac-tured bags can be initially tested to determineparticulate load (number and size distribution). Moredetailed studies can subsequently be performed tocharacterize the particulates, which can be used to de-termine their source. In this example, particulatecontrol and reduction strategies can be investigatedat two major levels: the final assembled product andthe components. Each can then be broken down usinga fishbone diagram to investigate ingress routes.Thisdetailed analysis can reveal areas where controls andreduction can be applied:

Final product assembly

• 100% visual inspection of the final containers• Particulate testing as part of product validation• Some level of product lot testing for particu-

lates (depending on volume)• General particulate characterization for source

identification• Regular clean room monitoring and acceptance

levels• Cleaning practices for clean room, machines and

incoming raw materials/components• Evaluation and maintenance of gowning

Component suppliers of final product

• Audit suppliers on a regular basis• Inquire about component particulate testing• Manufacturing environment (class or grade)• Type of packaging and packaging environment• Control and mitigation procedures• Personnel and gowning procedures/requirements• If component is “dirty,” can same be sourced else-

where (secondary source)?• Cleaning procedure (not ideal, as validation is

likely required)

Numerous steps can be considered and put intopractice for each product—whether it is a single-usecomponent, raw material or the final cell therapyproduct—but each requires a detailed analysis. Evenif some of the steps described here are addressed andcontrolled appropriately, the overall particulate loadfor a given cell therapy product can still be reduced,as illustrated in Figure 3. Understanding the originof particulates and their general impact on the productand the intended patient is necessary, and taking actionsto control and reduce particulates will be key to con-tinuous improvement. Because time and cost will bothbe significantly affected, cell therapy suppliers andsponsors need to consider the intended application ofthe products to determine the potential particulate risk.

Concluding remarks

Given the nature of the processes and products in celltherapy, particulates are common and likely unavoid-able at some level. Currently there are no limits orstandards specific to particulates and cell therapy prod-ucts. On the basis of the information described in thisreview, a number of factors are important (size,number, source and infusion route) when determin-ing possible risks, which include not only risks to theintended patient population but also risks to the celltherapy product itself. As demonstrated, particulatescan be added during each manipulation and can con-tinue to accumulate throughout the entire cell therapymanufacturing process. The first step is to know theparticulate composition of the final product.With thisinformation, risk can be addressed, and subsequentparticulate controls and mitigation actions can beimplemented as needed by both cell therapy spon-sors and suppliers.

Although following current guidance and stan-dards developed for other industries is commonpractice, we propose that the unique nature of celltherapy requires its own risk-based approach.The celltherapy industry needs to maintain its awareness andcontinue monitoring and controlling particulates, basedon the general understanding of particulates and the

ARTICLE IN PRESS

12 D. Clarke et al.

literature documenting patient risk.This understand-ing and the supporting data will play a key role indetermining suitable standards as the industry matures,ensuring the highest patient safety and maintainingthe availability of cell therapies to treat illness and life-threatening disease.

Disclosure of interests:The authors have no com-mercial, proprietary, or financial interest in the productsor companies described in this article. The authorsalone are responsible for the content and writing ofthe paper.

References

[1] Alliance for Regenerative Medicine. Alliance for regenerativemedicine quarterly data report Q3, <www.alliancerm.org>;2015 [accessed 12.15].

[2] Negrin RS. Introduction to the review series on “Advancesin hematopoietic cell transplantation”. Blood 2014;124(3):307.

[3] Das TK. Protein particulate detection issue in biotherapeuticdevelopment—current status. AAPS PharmSciTech 2012;13(2):732–46.

[4] Langille SE. Particulate matter in injectable drug products.PDA J Pharm Sci Technol 2013;67(3):186–200.

[5] Davis K, Evrard P, Isberg E, Jenness E, Johnson M, KollarC, et al. Recommendations for testing, evaluation, and controlof particulates from single-use process equipment. Washington,DC: Bio-Process Systems Alliance; 2014 <www.bpsalliance.org>.

[6] ISPE. ISPE good practice guide: assessing the particulatecontainments performance of pharmaceutical equipment.Tampa, FL: International Society for PharmaceuticalEngineering; 2012 <www.ispe.org>.

[7] Clarke D, Harati D, Martin J, Rowley J, Keller J, McCamanM, et al. Managing particulates in cellular therapy. Cytotherapy2012;14(9):1032–40.

[8] Cunha B, Aguiar T, Silva M, Silva R, Sousa M, Pineda E,et al. Exploring continuous and integrated strategies for theup- and downstream processing of human mesenchymal stemcells. J Biotechnol 2015;10(213):97–108.

[9] Bukofzer S, Ayres J, Chavez A, Devera M, Miller J, Ross D,et al. Industry perspective on the medical risk of visibleparticles in injectable drug products. PDA J Pharm SciTechnol2015;69(1):123–39.

[10] Fischer UM, Harting MT, Jimenez F, Monzon-Pasadas WO,Xue H, Savitz SI, et al. Pulmonary passage is a major obstaclefor intravenous stem cell delivery: the pulmonary first-passeffect. Stem Cells Dev 2009;18(5):683–91.

[11] Kringsholm B, Christofferson P. The nature and theoccurrence of birefringent material in different organs in fataldrug addiction. Forensic Sci Int 1987;34(1–2):53–62.

[12] Coughlin PA, Mayor AD. Arterial consequences of recreationaldrug use. Eur J Vasc Endovasc Surg 2006;32(4):389–96.

[13] Reedy JS, Kuhlman JE, Voytovich M. Microvascularpulmonary emboli secondary to precipitated crystals in apatient receiving total parenteral nutrition: a case report anddescription of the high-resolution CT findings. Chest1999;115(3):892–5.

[14] Pesko LJ. Physiological consequences of injected particulates.In: Knapp JZ, Barber TA, Lieberman A, editors. Liquid- andsurface-borne particle measurement handbook. New York:Marcel Dekker; 1996. p. 661–86.

[15] Barber TA. Control of particulate matter contamination inhealthcare manufacturing. Denver, CO: Interpharm Press;2000.

[16] Jack T, Brent BE, Boehne M, Muller M, Sewald K, BraunA, et al. Analysis of particulate contamination of infusionsolutions in a pediatric intensive care unit. Intensive Care Med2010;36(4):707–11.

[17] Tawde SA. Particulate matter in injectables: main cause forrecalls. J Pharmacovigil 2014;3(1):1–2.

[18] Muylle L. The role of cytokines in blood transfusion reactions.Blood Rev 1995;9(2):77–83.

[19] Jaganathan H, Godin B. Biocompatibility assessment ofSi-based nano- and micro-particles. Adv Drug Deliv Rev2012;64(15):1800–19.

[20] Moerke C, Mueller P, Nebe B. Attempted caveolae-mediatedphagocytosis of surface-fixed micro-pillars by humanosteoblasts. Biomaterials 2016;76:102–14.

[21] Oren R, Farnham AE, Saito K, Milofsky E, Karnovsky ML.Metabolic patterns in three types of phagocytizing cells. J CellBiol 1963;17(3):487–501.

[22] Laquerriere P, Grandjean-Laquerriere A, Guenounou M,Laurent-Maquin D, Frayssinet P, Nardin M. Correlationbetween sintering temperature of hydroxyapatite particles andthe production of inflammatory cytokines by humanmonocytes. Colloids Surf B Biointerfaces 2003;30(3):207–13.

[23] Foged C, Brodin B, Frokjaer S, Sundblad A. Particle size andsurface charge affect particle uptake by human dendritic cellsin an in vitro model. Int J Pharm 2005;298(2):315–22.

[24] USP <1>. Injections. U.S. Pharmacopeia National Formulary;2009.

[25] EP 2.9.20. Particulate Contamination:Visible Particles. EuropeanPharmacopeia; 2008.

[26] JP 6.06. Foreign Insoluble Matter Test for Injections. JapanesePharmacopoeia; 2013.

[27] USP <788>. Particulate Matter for Injections. U.S. PharmacopeiaNational Formulary; 2012.

[28] ASTM F24-09 (2015). Standard test method for measuringand counting particulate contamination on surfaces. WestConshohocken, PA: ASTM International; 2015 <www.astm.org>.

[29] Thornton RM. Pharmaceutical sterile cleanrooms. PharmTech1990;14:44–8.

[30] Austin PR. The encyclopedia of cleanrooms, bio-cleanroomsand aseptic areas. 3rd ed. Livonia, MI: Acorn Industries; 2000.

[31] Mollah H, Baseman H, Long M, editors. Risk managementapplications in pharmaceutical and biopharmaceuticalmanufacturing. Hoboken, NJ: JohnWiley & Sons, Inc.; 2013.

Appendix: Supplementary material

Supplementary data to this article can be found onlineat doi:10.1016/j.jcyt.2016.05.011.

Appendix

Case study

Product P is a cell therapy product that is manufac-tured from a bone marrow donation. The product ismanufactured in the following steps using the follow-ing components:

• Collection of 30 mL bone marrow aspirate intoa sterile blood collection bag

ARTICLE IN PRESS

Managing particulates in cell therapy 13

• Isolation of adherent cells (tubes, plastictubes)

• Culturing cells for four passages (tissue cultureflasks/multi-trays, tubes, and buffer/mediabags)

• Cryo-banking of the cells (2-mL cryovials)• Thawing the cells and passaging in a bioreactor

for additional three weeks (medium bags, tubes,carriers)

• Downstream processing (centrifuge kit, tubes,buffer bags)

• Final formulation (spinner flask, tubes, bags,containers)

• Cryopreservation (vial or cryobag)

The suggested approach taken in this case study,based on the aforementioned strategy, was divided intofour main steps:

1. Identification of the source and nature of theparticulates

2. Risk analysis and risk ranking3. Mitigation, including defining measures to reduce

and control the amount of particulates4. Long-term trend analysis and control

Step 1: Identification of the particulate sourceand natureTo understand the baseline amount and nature of par-ticulates in the raw materials and disposables, threefull kits including all raw materials, tubes and dispos-ables were prepared. Filtered water (0.22 μm) in a cleanand closed container was used to fill the disposables.The water from each disposable was carefully testedto see whether particulates could be observed, quan-tified and identified.This initial step was used to mapthe source and baseline amount of particulates thatdo not originate from the culture.

Step 2: Risk assessmentThe disposables were ranked by two main categories:their step in the process and the quantity/nature of par-ticulates found (Supplemental Table SI). The closerthe process step in which the disposable was used wasto the final formulation, the higher the risk rank it re-ceived. Additionally, as volumes change, a score forthe impact was determined. This impact score tookinto account the total number of particulates that thedisposable introduced to the process; for example, amedia bag of 10 L with 10 particulates received a lowerscore then a 1-L buffer bag that introduced 15 par-ticulates, but the 10-L bag received the same score asa 20-L bag with the same amount of particulates becauseof the introduction of the same total number of par-ticulates to the product.Additionally, the nature of theparticulate and its safety was assessed and scored.

Step 3:MitigationFollowing the risk assessment, options for mitiga-tion were then established. Examples are shown inSupplemental Table SII.

Step 4: Long-term trend analysis and controlTo ensure the quality of each batch, a trend analysisof particulates identified during the visual inspec-tion of the process steps must be logged and analyzed.Every new batch of material that is introduced to themanufacturing process must be initially tested forthe base level of particulates in the material, similarto the study described here; this can also determinea maximum particulate level.

It is important to work with suppliers to set cri-teria that are achievable and acceptable to both parties.Establishing criteria that are too low or to which a sup-plier cannot commit must be defined, and further in-house mitigation steps (such as filtration) should beintroduced.

ARTICLE IN PRESS

14 D. Clarke et al.