In the meantime, techniques for the culture of pure normal mammary epithelium from adult mice were developed (14). A more or less prolonged digestion with collagenases separates the epithelium from the adipose tissue in which it is normally embedded. In prelactating mammary glands, an extremely well-developed epithelial system replaces almost completely the connective tissue usually abundant in a resting organ. Explantation of this material gives large monolayered sheets of thin epithelial cells in pure colonies (fig. 2).

With a procedure identical to the one devised with organ cultures, C57BL mammary epithelium grown in roller tubes receives the agent in the form of RIII milk diluted to 10^-3 in the culture medium. Transfers are then made routinely once a month, in 1000-fold dilutions over a period of 1 year. The results of systematic bioassays on groups of 20 mice (table 2) show that only 2 tumors are induced after 1 month of cultivation, 2 after the second, and 1 after the third. It therefore appears that, though the milk agent can be maintained for at least 3 months in this type of culture, it does not multiply and is eventually eliminated.

Table 2. Tumor Induction With Roller-Tube Culture Extracts*

Dilutions of primary inoculum

Months of cultivation

RT+ MTA**

Controls, RT

10^-3

1

2

0

10^-6

2

2

0

10^-9

3

1

0

10^-12

4

0

0

10^-15

5

0

0

10^-18

6

0

0

10^-21

7

0

0

10^-24

8

0

0

10^-27

9

0

0

10^-30

10

0

0

10^-33

11

0

0

10^-36

12

0

0

*Twenty mice were Inoculated each time both with the experimental and control media. **RT: roller-tube extract; MTA: mammary-tumor agent.

Two possible alternatives can be considered to explain such results: (1) Instead of from embryonic tissues, the roller-tube cultures are derived from adult mammary glands in the prelactation stage. It has been shown by Andervont (15) and Dmochowski (16) that a natural resistance to the agent develops in older mice. This is particularly true of the C57BL mice not genetically inclined to transmit the milk factor. (2) The second alternative is the difference in complexity between organotypic and conventional homogeneous cultures. Could it be possible that the multiplication of the factor is dependent on specific metabolites supplied by the surrounding tissues?

Embryonic Tissues

The first consideration is easy to test. Since the mammary epithelium differentiates in organ cultures from the epidermis, whole skin from 10- to 12-day embryos should provide a good tissue substrate for virus multiplication. Proceeding in the same manner as with the adult mammary epithelium, we made cultures of embryonic skin and inoculated them with RIII milk diluted to 10^-3 in the medium. The infectivity was tested by bioassays through 4 successive passages and 1000-fold dilutions. Three tumors could be induced with the culture medium after

3 days of incubation and only one after the first passage (21 days). From a total of 130 experimental mice it is evident that the embryonic quality of the tissue is not alone sufficient to insure the proliferation of the agent.