Human mesenchymal stem cells enhance the systemic effects of radiotherapy.

The outcome of radiotherapy treatment might be further improved by a better understanding of individual variations in tumor radiosensitivity and normal tissue reactions, including the bystander effect. For many tumors, however, a definitive cure cannot be achieved, despite the availablity of more and more effective cancer treatments. Therefore, any improvement in the efficacy of radiotherapy will undoubtedly benefit a significant number of patients. Many experimental studies measure a bystander component of tumor cell death after radiotherapy, which highlights the importance of confirming these observations in a preclinical situation. Mesenchymal stem cells (MSCs) have been investigated for use in the treatment of cancers as they are able to both preferentially home onto tumors and become incorporated into their stroma. This process increases after radiation therapy. In our study we show that in vitro MSCs, when activated with a low dose of radiation, are a source of anti-tumor cytokines that decrease the proliferative activity of tumor cells, producing a potent cytotoxic synergistic effect on tumor cells. In vivo administration of unirradiated mesenchymal cells together with radiation leads to an increased efficacy of radiotherapy, thus leading to an enhancement of short and long range bystander effects on primary-irradiated tumors and distant-non-irradiated tumors. Our experiments indicate an increased cell loss rate and the decrease in the tumor cell proliferation activity as the major mechanisms underlying the delayed tumor growth and are a strong indicator of the synergistic effect between RT and MSC when they are applied together for tumor treatment in this model.


Dosimetric evaluation
The aim of this study has been firstly, to measure the absorbed dose on the treated tumor with extreme precision and secondly, to estimate what were the levels of the received dose in the tumors that had to be protected from the radiation action using the shield designed for evaluating the bystander effect in vivo.
Ionizing radiation was delivered by X-Ray TUBE (YXLON, model Y, Tu 320-D03) using a voltage of 240.4kV, a working current of 13.0 mA, a 0.32 mm Cu filter system, a 5 mm diameter focus of irradiation, a target distance of 10 cm and an irradiation field of 0.78 cm 2 , and a dose rate of 1502 ± 0.3 mGy/min.
For the radiochromic film dosimetry system we have used the Gafchromic film EBT3 [1, 2] and a scanner Epson Expression 10000XL. We have applied two different assemblies, which are described in the following paragraphs.
In the first we reproduced the irradiation conditions in which is it possible to quantify the absorbed dose, taking into account the international protocols [3]. The films are irradiated in whole backscattering conditions on a water phantom at 50 cm of the X-Ray source. This allows us to verify the values of the absorbed doses using the ionometric system normally used in the calibration of X-Rays beams of medium and low energy.
The second assembly seeks to reproduce the irradiation conditions of the mice (Fig. 1).
In this case the films are placed between two sheets of RW3 equivalent-water material (PTW, Friburgo, Germany), 160mm 160mm 10mm and 160mm  160mm  2mm. Figure 1: Assembly that mirrors the irradiation conditions of the animals (mice phantom). Film A is approximately 2.3 mm thick and the width of film B is 0.28 mm.
24 hours after the phantom irradiation we performed the scanning of the films and their analysis using ImageJ software.
Both the irradiated and the unirradiated films (control film) are scanned in the transparence mode. All the film scanning was performed in the same direction and at the same region of the scanner. We have applied the color method (48 bit), the green channel and a resolution of 96 ppp. The pixel values associated to every position are transformed to optical densities with the help of the equation: where net OD is the net optical density VP control the calculated pixel mean value for the control film after defining an area of 2x2 cm, and VP exposed the pixel values corresponding to each position in the irradiated films. To obtain the relative dose is it necessary to perform another mathematical transformation using the following expression [4]: Although the RW3 material for dosimetry media and low energy photon beams is not recommended, it is important to note that the differences observed between water and RW3 for energy and selected depth (2 mm) are not significant [4].

Film irradiation in reference conditions
Following the method previously described we have irradiated a film fragment at the nominal doses of 0 The mean difference between nominal doses and established doses with the film is 4%, which shows a good dosimetric concordance, taking into account the uncertainty bonded to the film.

Film irradiation in the conditions of animal treatment
We have irradiated three pairs of films placed in the positions indicated in Figure   To sum up we can state that the absorbed dose in the tumor is approximately, 2 Gy, whereas the dose received in the contra-lateral tumor is approximately 0.1 Gy.

Quantitative Real-Time PCR
This method allowed us to access the optical density values corresponding to TRAIL and DKK3 protein content in the cells. The human-specific primers used for relative We used RPLP0 (fw: CAGATTGGCTACCCAACTGTT; rev: GGCCAGGACTCGTTTGTACC) mRNA levels as internal reference gene to normalize the data.

ELISA of umbilical cord MSC whole-cell assay
Images of the wells were treated with the "Threshold colour" tool to select only colonies larger than 50 cells and then transformed to binary to allow particle analysis. The colony number and size was determined for each experiment measuring all colonies which were bigger than 50 cells. To take this into consideration we used the "Total Area" parameter returned by the ImageJ's Particle Analysis tool and the total area occupied by the colonies in each experimental condition was compared with the colonies from the untreated condition. Briefly, the cells in exponential growth were harvested, counted, and seeded in an appropriate number on 24-well plates. Cells were left for 24 hours to adhere to the culture flasks and then fixed in 4% PFA for 10 minutes at room-temperature, washed with PBS three times and stored at 4 ºC until used. The cells were then permeabilized by incubating them with PBS with 0.1 % saponin at room temperature for 10 minutes, washed three times with PBS with 0,05% Tween-20 (PBSt) and then incubated with PBS with 1% BSA for 1 hour at room temperature to block unspecific epitopes. After washing thoroughly with PBSt, cells were incubated with a biotinilated antibody against TRAIL (BD, 550948c) or against DKK3 (R&D, 842226) mixed with the steptavidin conjugated to horseradish-peroxidase for 1 hour at room temperature. Cells were then washed at least 7 times with PBSt, incubated at room temperature, protected from light for 30 minutes with the freshly prepared substrate solution (BD OptEIA™, 555214). The reaction was stopped with phosphoric acid 1 M.
Samples were monitored at 450 nm with an automatic plate reader (TitertekMultiskan plus, ICN Flow).

Tumor volume calculation
Tumor volume was calculated using the formula ( ) ( ), a and b being the values for the smaller and the larger diameter, respectively. Mice bearing tumors larger than 64 mm 3 were randomly distributed into 4 different groups: control, radiotherapy, MSC therapy and radiotherapy plus MSC therapy.

Histopathological and immunohistochemical studies
Tumors from a human melanoma cell line were xenografted in 32 mice (8 from each study group) and were immediately fixed in 10% buffered formalin for 48 h, and then embedded in paraffin, and 4 µm sections were dewaxed, hydrated, and stained using the hematoxylin-eosin technique. On these slides we determined the mitotic index, the necrotic areas and apoptotic cells observed outside the necrotic fields. Moreover, a complete microscopic study of pelvic, abdominal and thoracic organs was done to assess possible metastasis. For further details see supplementary materials.
The mitotic index was determined by counting the number of mitotic figures under microscopic observation in 10 high power fields (40x objective) from each of the 8 mice from each group, following the procedure previously described [Evans A.T. 1992].
Data were represented as mean±SEM for each of the groups studied (Fig. 7). Besides the mitotic index determination, data of the tumors from the whole section were gathered on the semi-quantitative score (0: absent, 1: low, 2: medium, and 3: intense) to identify the necrosis zones observed, and the number of apoptotic cells placed outside the necrotic fields. We assessed 10 randomly selected microscopic fields obtained from