Prospective use of the single-mouse experimental design for the evaluation of PLX038A

Abstract

Purpose: Defining robust criteria for drug activity in preclinical studies allows for fewer animals per treatment group, and potentially allows for inclusion of additional cancer models that more accurately represent genetic diversity and, potentially, allows for tumor sensitivity biomarker identification.

Methods: Using a single-mouse design, 32 pediatric xenograft tumor models representing diverse pediatric cancer types [Ewing sarcoma (9), brain (4), rhabdomyosarcoma (10), Wilms tumor (4), and non-CNS rhabdoid tumors (5)] were evaluated for response to a single administration of pegylated-SN38 (PLX038A), a controlled-release PEGylated formulation of SN-38. Endpoints measured were percent tumor regression, and event-free survival (EFS). The correlation between response to PLX038A was compared to that for ten models treated with irinotecan (2.5 mg/kg × 5 days × 2 cycles), using a traditional design (10 mice/group). Correlations between tumor sensitivity, genetic mutations and gene expression were sought. Models showing no disease at week 20 were categorized as 'extreme responders' to PLX038A, whereas those with EFS less than 5 weeks were categorized as 'resistant'.

Results: The activity of PLX038A was evaluable in 31/32 models. PLX038A induced > 50% volume regressions in 25 models (78%). Initial tumor volume regression correlated only modestly with EFS (r² = 0.238), but sensitivity to PLX038A was better correlated with response to irinotecan when one tumor hypersensitive to PLX038A was omitted (r² = 0.6844). Mutations in 53BP1 were observed in three of six sensitive tumor models compared to none in resistant models (n = 6).

Conclusions: This study demonstrates the feasibility of using a single-mouse design for assessing the antitumor activity of an agent, while encompassing greater genetic diversity representative of childhood cancers. PLX038A was highly active in most xenograft models, and tumor sensitivity to PLX038A was correlated with sensitivity to irinotecan, validating the single-mouse design in identifying agents with the same mechanism of action. Biomarkers that correlated with model sensitivity included wild-type TP53, or mutant TP53 but with a mutation in 53BP1, thus a defect in DNA damage response. These results support the value of the single-mouse experimental design.

Keywords: Alternative experimental design; Drug evaluation; Genetic diversity; Pediatric cancer; Response criteria; Single mouse study.

Fig. 1

Responses of 31 individual xenograft models in mice treated with a single administration of PLX038A. Tumor diameters were measured weekly until tumor reached endpoint criteria (4-fold volume at day of treatment) or day 140 when experiments were terminated.

Fig.2

A, ‘Waterfall’ plot showing the maximal tumor volume regression (100% is a volume below level of detection ~40 mm³) in mice treated with a single administration of PLX038A; B, “Swimmer’ plot showing the EFS for 31 xenograft models in treated xenografts. Tumors marked with an asterix are those where there was no detectable tumor at termination of the experiment at day 140. C, The relationship between the maximal tumor volume reduction and EFS for 31 xenograft models in mice treated with PLX038A (R² = 0.2293). Arrows mark overlapping coordinates for SK-NEP-1 and RBD-2 [day 38] and Rh36, EW-8, TC-71, KT5 and KT11 [day 140]).

Fig. 3

A, The relationship between EFS determined after irinotecan (2.5 mg/kg daily × 5; EFS(IRN)) and EFS following a single administration of PLX038A (EFS(PLX038A). KT11 xenografts are identified as an outlier sensitive to PLX038A. Data for irinotecan are median EFS for groups of 10 treated mice; B, The same data without KT11 (R² = 0.6844).

Ƒig. 4.

Differentially mutated genes between resistant and sensitive tumors. Solid black cell indicates a non-synonymous mutation. The three genes shown in the heatmap have the most significant p values (p=0.18) from Fisher’s exact test without reaching statistical significance.