Perspective: Recommendations for benchmarking pre-clinical studies of nanomedicines

Abstract

Nanoparticle-based delivery systems provide new opportunities to overcome the limitations associated with traditional small molecule drug therapy for cancer, and to achieve both therapeutic and diagnostic functions in the same platform. Pre-clinical trials are generally designed to assess therapeutic potential and not to optimize the design of the delivery platform. Consequently, progress in developing design rules for cancer nanomedicines has been slow, hindering progress in the field. Despite the large number of pre-clinical trials, several factors restrict comparison and benchmarking of different platforms, including variability in experimental design, reporting of results, and the lack of quantitative data. To solve this problem, we review the variables involved in the design of pre-clinical trials and propose a protocol for benchmarking that we recommend be included in in vivo pre-clinical studies of drug delivery platforms for cancer therapy. This strategy will contribute to building the scientific knowledge base that enables development of design rules and accelerates the translation of new technologies.

Introduction

Chemotherapeutics are inherently toxic and hence treatment is invariably a balance between inducing cancer cell death and minimizing the adverse side effects associated with drug accumulation in normal tissue and organs (1,2). Nanotechnology provides the possibility of creating delivery systems where the design constraints are decoupled, allowing new approaches for reducing the unwanted side effects of systemic delivery, increasing tumor accumulation, and improving efficacy (3–5). The development and translation of new nanomedicines is extremely challenging and hence progress in the field has been relatively slow. Of the 6 FDA-approved nanomedicines for cancer therapy, two are antibody-drug conjugates, 3 are liposomal formations of anti-cancer drugs, and one is a protein-drug complex (6). Progress has also been hindered by the largely empirical approach to research and development (7).

Building a knowledge base is key to advancing a scientific field since it enables the development of design rules that, in turn, accelerate the translation of new technologies. Pre-clinical trials of nanomedicines are usually designed to assess efficacy by measuring changes in tumor size and/or survival rates following administration in an animal model. However, to develop design rules for drug delivery platforms it is essential to understand how the physico-chemical properties (e.g. size, shape, composition, zeta-potential, surface chemistry, etc.) influence pharmacokinetic parameters (area under the curve, clearance rate, distribution volume, and clearance half-time), biodistribution (uptake in organs), tumor accumulation, tumor size (growth rate), and survival (Kaplan - Meier curves).

Despite the large number of pre-clinical trials of drug delivery platforms, surprisingly few report quantitative data that that would be useful in developing platform design rules. In a recent meta-study we reviewed 74 pre-clinical trials of drug delivery platforms that reported quantitative values of tumor accumulation (7) and showed how the variability in experimental design significantly limits the ability to make the comparisons necessary to develop design rules. Specific problems include: differences in the animal model, inconsistent reporting of tumor size/weight, inconsistent reporting of dose, inconsistent reporting of physico-chemical properties, tumor accumulation reported at different time points, and variability in controls (especially for active targeting).

The ability to make comparisons between studies with different platforms is key to accelerating progress in the field and increasing the impact of the investment in research. To overcome this problem we propose a standard protocol that could be included in pre-clinical trials to allow benchmarking between studies. In vivo pre-clinical trials are time consuming and often expensive, and hence recommendations for benchmarking should be convenient, limited to a small number of animals, and be limited to a small number of measured parameters.

Here we consider the variables involved in the design of pre-clinical trials, including animal model, tumor size, dose, and measured variables such as physico-chemical properties and pharmacokinetics/biodistribution. We propose a protocol for benchmarking that could be included in in vivo pre-clinical studies of drug delivery platforms for cancer therapy (Table 1). Briefly, we propose that pre-clinical trials of drug delivery platforms should include measurement of: (1) the concentration of the delivery vehicle in blood, and (2) tumor accumulation in athymic Nu/Nu mice with subcutaneously implanted LS174T cells grown to a tumor size of 8 – 10 mm in diameter. Experiments should be performed at a minimum of three fixed time points, at 6 h, 24 h, and 48 h post-injection. Tumor accumulation should be obtained using a suitable quantitative method and reported as %ID and %ID/g. We recommend that experiments are performed at a dose of 10¹³ particles per mouse (approximately 20 g). These additional experiments will typically require 21 mice (typically 7 at each time point). Data from such experiments will allow direct comparison and benchmarking of different platforms, and contribute to the development of design rules that will allow more efficient utilization of research resources and accelerate translation to the clinic.

TABLE 1.

Protocol for benchmarking drug delivery platforms designed to treat solid tumors. Standardization of experimental procedures and corresponding reporting metrics will enable direct comparison and contribute to the development of design rules that will accelerate progress in the field.

	Guideline	Reporting
animal model	LS 174T subcutaneous xenografts grown to 8–10 mm in athymic nu/nu mouse	weight (g), diameter (mm)
physicochemical properties	size, shape, composition, surface chemistry, zeta potential collected	diameter (nm), shape, composition, surface chemistry, zeta potential
dose	1013 NP	number of NP
pharmacokinetics	6, 24, 48 h	% ID, mg of drug
tumor accumulation	6, 24, 48 h	% ID, % ID/g

Animal Models

Rodent models are widely used in cancer research for identification of carcinogens, the development of cancer therapies, and to study the molecular mechanisms of tumor growth and metastasis. Mouse models have been developed for primary tumors in most organs or organ systems, as well as models of metastasis, and include: environmentally-induced models, human tumor xenografts in immunocompromised mice, and genetically engineered mice (8,9). While mice are relatively inexpensive, they have a blood volume of only about 2 mL and hence it is difficult to extract sufficiently large blood samples for analysis of the concentration in blood at different time points from a single animal. Although the fraction of successful pre-clinical studies that survive clinical trials is very low (10), animal studies can provide insight into the relationship between physico-chemical properties and pharmacokinetic parameters. The predominant mouse models include spontaneous and environmentally-induced mouse models, immunocompromised mice for xenografting, and genetically engineered mice (see Supplementary Information for further details) with varying levels of physiological relevance and accessibility.

Recommendation for benchmarking

While GEM models and primary human tumor xenografts in humanized mouse models represent the state-of-the-art in animal models for cancer research they are technically challenging, expensive, and time consuming. These models can recapitulate many aspects of the tumor microenvironment associated with specific types of cancer, and hence are important in developing therapies and delivery strategies for different malignancies. Despite the limitations of subcutaneous xenografts using immortalized cell lines, these models are straightforward, relatively inexpensive, and tumors can be formed within a few weeks. Therefore, for benchmarking drug delivery platforms we recommend a subcutaneous xenograft in athymic Nu/Nu mice.

Cell line

A wide range of human and animal cell lines have been used for subcutaneous xenografts. Indeed a major problem in benchmarking drug delivery platforms is the wide range of cell lines used for subcutaneous xenografts (7). In a recent meta study of nanomedicines that reported tumor accumulation, 68/74 studies used xenograft mouse models (7). These 68 studies used 35 different cell lines to form the xenografts, highlighting the variability in experimental design. Several commonly used cell lines are the 4T1, C26, KB, and LS174T lines (7) (see Supplementary Information for details).

Recommendation for benchmarking

We propose subcutaneous xenografts of 5 × 10⁶ LS174T cells injected as suspended cells in 50% growth media and 50% growth factor reduced Matrigel. LS174T implants grow reasonably quickly, reaching about 8 – 10 mm in size after 1.5 – 2 weeks, and do not appear to have a significant tendency to metastasize (11). Matrigel provides a viscous medium to minimize cell diffusion away from the injection site and decreases shape abnormalities during the initial tumor formation.

Tumor size

Tumor size is an important parameter in analysis of tumor accumulation when comparing nanomedicines. Small tumors, typically ≤ 1 mm³, are not well vascularized (12) and the normalized tumor accumulation (in terms of %ID/g) is expected to be less than in a well vascularized tumor. On the other hand, in large tumors the necrotic core may represent a significant fraction of the total tumor volume resulting in small values of accumulation compared to a tumor where the necrotic core is a small fraction of the tumor volume. The point at which tumors reach necrosis is highly dependent on their ability to self-vascularize. For details about tumor growth and the EPR effect, see the Supplementary Information.

In subcutaneous xenograft models, the time dependence of tumor size is widely used to assess efficacy. Tumor size is determined from measurement of the smallest and largest dimension of the tumor and a model of tumor shape (13). Tumor accumulation, when measured, is usually normalized by the tumor mass. While direct measurement of the mass of the resected tumor is preferred, in many cases tumor mass is determined from the tumor volume and an estimate of the tumor density. Both volume and density are subject to errors and should be avoided (13).

Recommendation for benchmarking

We recommend a tumors be grown to 8 – 10 mm in diameter or approximately 0.2 g in weight. This size range is typical for pre-clinical studies, and allows for a relatively large sample for analysis without a large degree of necrosis. The mass of the resected tumor should be reported along with the smallest and largest dimensions for calculation of size using the ellipsoid formula (13).

Physico-chemical evaluation

To establish design rules for drug delivery platforms it is essential to elucidate the relationship between physico-chemical properties and pharmacokinetics and tumor accumulation. Important design factors for physico-chemical evaluation include size, shape, surface chemistry, and aggregation state (14–16). All of these parameters are known to influence pharmacokinetics, biodistribution, uptake, and trafficking (17–22).

Recommendation for benchmarking

We recommend that all pre-clinical studies include characterization of size, shape, composition, surface chemistry, and zeta potential.

Dose

The dose of a drug is usually determined by the therapeutic index and hence standardization for pre-clinical trials is difficult. The dose of a drug delivery platform can be based on the amount of drug administered (e.g. mg of drug per kg of body weight or surface area), or can be reported as the number of nanoparticles administered. Reporting drug loading (number of molecules per carrier) is key to making appropriate comparisons. Similarly, dose reported as mg of the drug delivery platform per kg of body weight is only useful if the mass of a single nanoparticle and the drug loading are also reported. While radio-labeling is a commonly used method for imaging and determination of biodistribution, data are only useful for benchmarking if quantitative calibration curves are provided to convert radioactivity units (e.g. Curies or Bequerels) to dose.

There are six FDA-approved nanomedicines for solid tumor therapy: Doxil, Marqibo, Daunoxome, Abraxane, Brentuximab vedotin, and Trastuzumab emtansine (Table 2) (6). Doses for the liposomal drugs (Doxil, Marqibo, Daunoxome) vary from 2 – 93 nmol of nanoparticles for an average human patient (based on the mass of the drug). All three have a similar drug loading (about 10,000 molecules per liposome). The recommended dose for Abraxane is 100 – 260 mg m⁻² corresponding to 2 – 6 μmol of nanoparticles. The drug loading (≈ 10,000 per nanoparticle) is similar to the liposomal drugs. For the six nanomedicines, the number of particles per dose varies over about two orders of magnitude from about 10 nmol to 1 μmol.

TABLE 2.

Comparison of FDA-approved nanomedicines for targeting solid tumors and the corresponding free drugs. Drug loading is defined as the number of drug molecules per particle. Doses were obtained from clinical prescribing information and are based on the amount of the active drug. The number of nanoparticles administered per dose was determined from the drug loading and the dose. LD₅₀ data was obtained from the MSDS of the platform or free drug.

Drug	Size	Drug loading	Dose (drug)	Amount of platform	LD50
Doxil	100 nm	10,000 – 15,000	20 – 50 mg/m2	12 nmol	32 mg/kg
Daunoxome	60 nm	1,000 – 2,000	10 – 80 mg/m2	93 nmol per single use vial
Marqibo	100 nm	2000	2.25 mg/m2	2.16 nmol
Abraxane	130 nm	9	100 – 260 mg/m2	2 – 6 μmol	447 mg/kg
Brentuximab vedotin	2 nm	2 – 8	1.8 mg/kg	0.72 μmol
Trastuzumab emtansine	2 nm	2 – 8	3.6 mg/kg	1.5 μmol	0–5 mg/kg
Doxorubicin	580 Da	1	40 – 60 mg/m2	0.146 mmol	17 mg/kg
Daunorubicin	564 Da	1	30 – 60 mg/m2	96 μmol–0.19 mmol	20 mg/kg
Vincristine	923 Da	1	1.4 mg/m2	2.8 mmol	3 mg/kg
Paclitaxel	854 Da	1	100 – 175 mg/m2	0.19 – 0.35 mmol	12 mg/kg

Tumor accumulation and biodistribution are often determined using a radiolabel, and hence dose is reported in Curies or Bequerels. A disadvantage of this approach is that the absolute amount accumulated in the tumor can only be determined if the decay in emission of the radiolabel and the number of radiolabels per particle (for a given isotope) are reported.

To determine a suggested dose for benchmarking, we analyzed published quantitative studies of tumor accumulation in mouse studies in the literature (see Supplementary Information) (7). In these studies, the doses ranged from 10¹¹ to 10¹⁶ particles per mouse, with most around 10¹³ nanoparticles/mouse (or approximately 20 pmol/mouse). This dose, normalized to body weight, is within range of the doses of FDA-approved nanomedicines (Table 2).

Recommendation for benchmarking

We recommend a nanoparticle dose of 10¹³ particles/mouse. Dose should also be reported as mass of drug and mass of drug normalized to the mass of the animal.

Pharmacokinetics/Biodistribution

Pharmacokinetics and biodistribution are sometimes included in pre-clinical trials of nanomedicines. Pharmacokinetic parameters such as area under the curve, clearance rate, distribution volume, and elimination half-time provide important insight into stability in circulation, clearance by the kidneys and mononuclear phagocyte system (MPS), and the propensity for uptake in normal tissue. Biodistribution studies of the accumulation in organs, particularly liver and spleen, provide valuable information concerning possible toxicity.

For comprehensive comparison of drug delivery platforms it would be desirable to have detailed pharmacokinetics, biodistribution, and tumor accumulation data, along with tumor growth rate, and survival curves, under a given set of conditions. However, these time consuming and expensive experiments are unrealistic for benchmarking. As described previously, tumor accumulation is a parameter that is not widely reported but combines aspects of pharmacokinetics and extravasation by the EPR effect (see Supplementary Information). Therefore, tumor accumulation and concentration in circulation are key parameters for benchmarking.

In selecting time points for benchmarking, there are many competing factors. Circulation half-times for chemotherapeutics and FDA approved drug delivery platforms (Table 2) typically vary from 4 h – 4 days (6). Evidence suggests that the accumulation of a drug delivery system in a solid tumor increases with time following administration, and subsequently decreases at longer times (6,7). However, detailed measurements of concentration in circulation and tumor accumulation at multiple time points requires a large number of animals and hence is time consuming and expensive. Therefore, we recommend that blood concentration and tumor accumulation are measured at three time points: 6 h, 24 h, and 48 h.

Measurements at 6 h provide information on the concentration in circulation and tumor accumulation shortly after administration. This is particularly important for nanoparticle systems where a sustained concentration in blood is not required, for example for imaging. Measurements at 24 h provide insight into tumor accumulation at a later time point relevant for drug delivery systems where the concentration in circulation is likely to be considerably lower. Finally, at 48 h the concentration in circulation is generally very low resulting in a large concentration gradient for transport of unbound nanoparticles back into circulation. Therefore measurement of concentration in circulation and tumor accumulation at 48 h can provide insight into retention of the drug delivery system in the tumor. We emphasize that the goal is not to quantify pharmacokinetics or biodistribution, but to allow benchmarking between drug delivery systems at different time points.

Recommendation for benchmarking

We recommend that blood concentration and tumor accumulation are measured at 6 h, 24 h, and 48 h. A detailed discussion of methods for quantitative measurement is provided in Supplementary Information. We suggest that concentration in blood is reported as %ID and mg of drug. The tumor concentration should be reported as %ID and %ID/g tumor mass, from which the mass of the tumor can be inferred. Depending on the nanoparticle design and cargo, the analytical methods used to determine blood concentration and tumor accumulation may be very different. Therefore, appropriate controls, standards, and/or calibrations should be provided to ensure the validity of the reported metrics.

Suggested protocol for benchmarking pre-clinical studies of nanomedicines

Recipient animal and tumor model

Immunodeficient, 4 to 6 week old athymic nude female mice are housed in a specific pathogen-free animal facility in accordance with federal guidelines, including the Principles of Laboratory Animal Care, along with regulations from each institution’s Committee on the Use and Care of Animals.

Tumor cell line

The LS174T human colon carcinoma cell line is purchased from the American Type Culture Collection (Manassas, VA) and maintained in vitro at 37 °C, 5% CO₂ in Eagle’s Minimum Essential Medium with 10% fetal bovine serum and 1% penicillin/streptomycin. Cells should be used only below passage 15. Before injection in the mice, the cells are harvested with trypsin-EDTA solution, washed, and resuspended in PBS. The cell suspension (5 × 10⁶ cells in 50 μL of growth media and 50 μL of growth factor reduced Matrigel) is injected subcutaneously into one flank of each mouse. Tumors are allowed to grow for approximately 1.5 – 2 weeks until reaching a diameter of 8 – 10 mm. Tumor volume is then computed for the standard volume of an ellipsoid (13).

Drug administration

Animals are injected via lateral tail vein with less than 200 μL solution containing 10¹³ nanoparticles. Dose should be reported as the number of nanoparticles injected in addition to other relevant metrics, such as drug/radiolabel loading, total drug administered, etc.

Physico-chemical properties

The following data should be reported: average particle size, shape, composition, surface chemistry, and zeta potential.

Biodistribution and tumor accumulation of nanoparticle

At 6 h, 24 h, and 48 h the animals are euthanized and tumors are entirely resected and weighed. Each group includes a minimum of seven mice. Blood concentration should be reported as amount (mg) and %ID, and tumor concentration should be reported as amount (mg), %ID, and %ID/g tumor.

Conclusion and future perspectives

There has been rapid growth in the number of pre-clinical trials of drug and gene delivery systems for cancer therapy and imaging. However, these studies have not resulted in the development and validation of design rules to the extent that would be expected based on the number of publications. It is extremely challenging to move pre-clinical studies to clinical trials and ultimately to the clinic, as evidenced by the small number of FDA approved nanomedicines (6). However, the development of delivery systems remains largely empirical and the lack of standardization of pre-clinical studies has been a barrier to establishing robust design rules for nanoparticle development.

The ability to make quantitative comparisons between delivery platforms is key to elucidating the complex interactions that relate the properties of the delivery platform to efficacy. In general, it has been established that drug toxicity can be reduced by encapsulating the free drug (e.g. liposomes). Similarly, stability in circulation can be improved by developing strategies to minimize protein binding and evade the immune system. The efficiency of accumulation at a tumor site can be improved by active targeting of the delivery system or by increasing extravasation by the EPR effect. However, a more detailed understanding of the complex interactions of delivery platforms in the body would greatly accelerate progress in the field.