Patients whose tumors were implanted into mice were enrolled into a phase II clinical trial and treated with neoadjuvant docetaxel (75 mg/m2 i.v.) plus carboplatin (area under curve 6 i.v., every 3 weeks) for 6 cycles prior to surgery. The results of this clinical study (NCT 01560663) have been previously published (9). Clinical response was assessed by magnetic resonance imaging of the breast, prior and after chemotherapy. Pathological responses in surgical specimens were classified according to the residual cell burden (RCB) system by Symmans et al (10). The combination of docetaxel and carboplatin was highly active in women, achieving a 55% of RCB class 0 responses (complete clearance of the tumor from breast and axilla -pCR-) and a 13% of class I responses (minimal microscopic residual disease in breast without axillary residual tumor). Both classes of pathological responses are associated with a similar, excellent, survival outcome(11,12).
The clinical protocol was approved by the Ethical Human Being Committee of the participating hospitals. The IiSGM Animal Care and Use Committee and Comunidad de Madrid approved all the PDX procedures (PROEX 137/16). Procedures involving animal care complied with national and international laws and policies. The collection of samples was approved by the corresponding ethics committee (collection identifier: C.0000188 at ISCIII, Spain). Patients provided written informed consent before any procedure.
Mice and establishment of tumor xenografts
NOD.Cg-Prkdcscid IL2rgtm1Wjl (NSG) mice (female, 8-week-old) were obtained from Jackson Laboratories, Bar Harbor, ME, USA. The mice were maintained under specific pathogen-free conditions in the animal facility of one of the institutions (IiSGM). The mice were fed and provided with autoclaved water ad libitum.
Tumors were implanted into the 4thinguinal mammary fat pads (orthotopically) by two small incisions in the skin to expose the fat pad: one along the midline and one a short distance down the leg, to make a pocket into the center of the transplantation site. Tumor growth was monitored once per week by palpation and caliper measurement. Once the first tumor (named F0 in future references) reached an estimated volume of 1000mm3, it was removed for the expansion process (Figure 1).
Confirmation of the nature of TNBC xenografts
To confirm the human nature of implanted the tumors, an Alu probe in situ hybridization was carried out in all the 9 xenografts after engraftment (F0).
Serial 4-mm-thick sections were cut from each paraffin-embedded PDX tumor. Examinations of tumor morphology were conducted on slices processed for H&E staining.
The triple negative breast cancer status of the xenografts was established by immunohistochemistry The slides were incubated with the appropriate primary antibody as detailed: rabbit monoclonal anti-Estrogen Receptor Alpha (EP1, Ready to use, Agilent, IR084), mouse monoclonal anti-Progesterone Receptor (PgR636, 1/500, Agilent, M3569), rabbit polyclonal anti-c-erBb-2 (1/400, Agilent, A0485) and mouse monoclonal anti-KI-67 (MIB-1, Ready to use, Agilent, IR626). All these determinations were performance before and after treatment.
PDX treatment protocol
The same experiment was performed in all 9 PDX models. Xenografts from the second mouse to mouse passage (F2) were allowed to grow to an estimated volume of 200-250mm3, at which time mice were allocated into the following five groups of treatment with ten mice in each group: (a) vehicle (ethanol, 20.8mg/ml, 200ul); (b) Docetaxel 8.10 mg/kg i.v once a week(13,14); (c) Carboplatin 48.64 mg/kg i.p once a week; (15)(d) combination Docetaxel +Carboplatin 48.64 mg/kg+8.10mg/kg i.v+i.p once a week; (e) Doxorubicin2.5 mg/kg i.p once a week(16). Mice were treated for 28 days, monitored daily for signs of toxicity, and were weighed three times a week. (Figure 1)
Evaluation of response to drugs in PDX
Tumor size was evaluated three times per week by caliper measurements. Tumor volume was calculated using the following formula: Tumor Volume = [tumor length x tumor width2]/2.
Tumor growth inhibition (%TGI) was calculated as (1-(Average relative tumor volume
treatment group/Average relative tumor volume vehicle group) *100).
Experiments were terminated on day 35. The proportion of residual tumor cells and the volume of the residual tumors were measured for the 5 therapeutic groups of each xenograft.
Molecular characterization of breast cancer patient-derived xenografts
- Extraction RNA
The invasive area of the tumor from haematoxilin and eosin stained slides was delimited by a pathologist at the LAOT and was subsequently microdissected in 10 mm slices. RNA was then extracted using the FFPET RNA Isolation Kit (Qiagen). RNA concentration and quality were assessed with Nanodrop 2000 Spectophotometer (Thermo Scientific NanoDrop Products) according to A260/280 ratio, which needs to be around 2.00 for RNA.
- Extraction DNA from fresh tumour
DNA was extracted from fresh tumour sample of PDX using phenol/chloroform/isoamyl method. DNA concentration and quality were assessed with Nanodrop 2000 Spectophotometer (Thermo Scientific NanoDrop Products), according guidelines, concentration needs to be around 50ng/ul.
- Intrisic subtype classification
We analyzed the expression of the 50 genes included in the PAM50 assay and 5 additional housekeeper genes as described by Parker et al. (17). In the raw nCounter® transcript quantification the background was corrected using the negative probes and normalizing with their mean minus 2 standard deviations and those values were normalized by calculating the geometric mean of the 5 housekeeper genes. Subtype classification was based on the nearest of the 5 centroids.
- Fluorescence in situ Hybridization (FISH) of topoisomerase II alpha (TOP2A)
TOP2A amplification was measured by FISH (Vysis-Abbot). The probe used was a locus-specific identifier TOPO2A on 17q21-22, labeled in orange and the centromere (CEP17) was hybridized with spectrum green on 17p11.1-q11.1 (Vysis-Abbott). Sections 3µm thick were cut from each paraffin-embedded tumoral piece extracted from the 9 PDXs without treatment. All procedures for the FISH analysis were prepared according to the manufacturer’s instructions. Images were analyzed on a fluorescence microscope. TOP2A was determined in a minimum of 20 nuclei counted per case. A positive result was defined as TOP2A:CEP17 ratio of 2 or greater (18). A tumor was considered to have deleted TOP2A if the ratio was 0.8 or lower (19).
Sequencing analysis of PDX
Sequencing libraries were created using 25ng of genomic DNA from 45 tumor samples belonging to different experimental groups. The panel includes 111 genes exons of which 12 are specifically studied for hereditary breast cancer (BRCA1, BRCA2, PALB2, RAD50, RAD51C, RAD51D, BARD1, ATM, BRIP1, CHEK2, NBN, PTEN). The library preparation, target capture and sequencing were carried out in the genomic systems unit of the Gregorio Marañon Health Research Institute (IiSGM). The sequencing was performed using the MiSeq platform (Illumina), according to the manufacturer’s instructions.
For variant calling we focused on single nucleotide substitutions, small insertions and deletions (indels). The list of variants were filtered by rare variants (MAF<0.01) (I), variants in coding exons and splicing junctions (II), discarding variants predicted to produce synonymous amino acid changes without splicing alteration (we only selected high/medium impact variants: frameshift indels and non-sense, missense and canonical splice site variants) (III). Next, we discarded benign and probably benign classified variants (IV) based on different databases (CLINVAR: https://www.ncbi.nlm.nih.gov/clinvar/; Leiden open variation database: http://www.lovd.nl/3.0/home; - BRCA exchange: http://brcaexchange.org/; Locus specific databases: http://grenada.lumc.nl/LSDB_list/lsdbs; varsome (https://varsome.com/); Pubmed: https://www.ncbi.nlm.nih.gov/pubmed/). Variants were classified according ACMG guidelines. A minimum depth of 30% is required for the posterior validation analysis of pathogenic variants with Sanger Sequencing using the Integrative Genomics Viewer (IGV).
Statistical analysis and graphical presentations were performed using GraphPad Prism 5 and Microsoft Excel. Descriptive data were generally expressed as mean ±standard deviation or standard error of the mean. Statistical evaluations of the differences between groups were assessed using student’s t-test analyses, with a prespecified alpha of 0.05.