WAPiCre-mediated ablation of c-Myc in the mammary gland
To study the role of c-Myc in mammary gland development, we used a conditional approach, crossing c-mycfl/fl
] to WAPiCre
transgenic mice [35
]. The generated offspring will be referred to as wild type (WT, c-mycfl/fl
), heterozygous (c-mycfl/+
) and mutant (c-mycfl/fl
) mice. In animals positive for the WAPiCre
transgene, the complete open reading frame of c-myc
will be excised upon Cre expression (Figure ). To assess onset and extent of WAPiCre
expression, we performed immunohistochemistry (IHC) against Cre recombinase on sections from mutant mammary glands (Figure ). Cre expression was first detected at day 14.5 of pregnancy in scattered luminal alveolar cells. The number of Cre-expressing cells increased continuously until after parturition, when positive staining for Cre was seen in essentially all luminal cells. To monitor recombination, we performed polymerase chain reaction (PCR) on genomic DNA isolated from mammary glands at different developmental stages. The 220 base pair band, indicating the presence of the recombined allele, was first detected at day 14.5 of pregnancy (Figure ), consistent with the results from IHC. Starting then, levels of c-myc
mRNA decreased rapidly in glands of mutant mothers and were essentially undetectable throughout lactation (Figure ). With the commercially available antibodies, it has not been possible to detect c-Myc in the lactating mammary gland by IHC (data not shown; Klinakis et al
]). Since the half-life of c-Myc protein and mRNA is short [37
], it is likely that mutant glands have little or no c-Myc by the onset of lactation. Finally, mRNA levels of the cell cycle inhibitor p21Cip1
, a well-studied target of c-Myc-mediated repression [38
], were upregulated in c-Myc-deficient glands during lactation (Figure ), which is in agreement with the functional loss of c-Myc in mutant glands.
Figure 1 Targeted disruption of c-Myc in the mammary gland. (a) Schematic diagram of c-myc floxed allele and recombined allele after Cre-mediated excision of floxed region. The position of the 220 base pair (bp) polymerase chain reaction (PCR) product for detecting (more ...)
c-Myc mutant mothers display a lactation defect with less efficient milk production
Monitoring survival and weight of newborn pups is routinely used as a measure of lactation [40
]. Thus, we performed a pup weight analysis to examine the efficiency of milk production in WT and mutant females. Growth curves generated from seven foster pups per mother showed that pups nursed by mutant mothers grew significantly slower compared with pups nursed by a WT mother (Figure , left panel). However, when comparing a mutant mother nursing only two foster pups to a WT mother nursing six pups, there was no significant difference in pup body weight (Figure , right panel). This suggests that milk quantity, but not quality might be affected in c-Myc-deficient glands.
Figure 2 Ablation of c-Myc in mammary glands results in impaired lactation due to reduced milk volume. (a) Growth analysis of pups nursed by wild type (WT) or mutant mothers. Data are shown as average body weight plus standard deviation. Left panel: analysis of (more ...)
To test this hypothesis, we first examined milk composition. Milk samples taken from WT and mutant mothers at day 14.5 and 15.5 of lactation were analyzed for protein, lactose and fat content, the three major milk components. On a Coomassie stained gel, milk protein pattern and concentration were identical in equal volumes of milk from WT and mutant mothers (Figure , caseins are indicated) (see also Marte et al
]). Furthermore, the concentration of lactose, the major carbohydrate and osmole in milk, as well as the fat content were determined in milk samples from a group of five animals made up of WT, heterozygous (showing no overt phenotype) and mutant mothers (Figure and ). Lactose concentration was determined in a colorimetric assay on skim milk samples, whereas fat content was measured as the ratio of cream layer length over total milk length after centrifugation ('creamatocrit') (Lucas et al
]). While one heterozygous mother showed a slightly decreased lactose concentration, likely due to natural variation (Figure ), there were no consistent alterations in either lactose or fat content within the samples.
Next, to compare the approximate amount of milk produced in the lactating glands from WT and mutant mothers, the following experiment was performed. In the first setting, mothers were sacrificed immediately after removing them from their actively suckling pups. In the second setting, mothers were removed from their pups and sacrificed 2 hours later, which allows the glands to fill with milk. When comparing high magnifications of whole mount preparations taken from actively nursing mothers, glands from WT and mutant mice looked nearly identical (Figure , panels a and b). However, only the WT females showed clear signs of milk-filling, displaying large, distended alvoeli after 2 hours without pups (Figure , panel c, arrows), while glands of mutant mothers appeared only slightly distended (Figure , panel d).
Finally, we examined the milk proteins via a Western analysis carried out on protein lysates made from lactating mammary glands of WT and mutant mothers. Equal amounts of protein were loaded and membranes probed with a rabbit anti-milk serum [41
], producing a staining pattern of multiple milk proteins (Figure ). The blot shows that mutant protein lysates contain less milk protein than the corresponding WT lysate at day 5.5, 10.5, and 15.5 of lactation. The level of α-tubulin, used as a loading control, was the same in each paired WT and mutant sample. Taken together, these results clearly suggest that the reduced nursing ability in c-Myc mutant mothers is due to decreased or slower milk production, while milk composition is essentially the same in mutant and WT mothers.
Alterations in alveolar density and secretory activity
To analyze the lactation defect in more detail, we investigated the morphology of glands from lactating, actively nursing WT and mutant mothers via IHC. Cytokeratin (CK) 18-stained cross-sections were scanned (Figure ), revealing that mutant glands contained more unstained stromal area than WT glands. To obtain quantitative results, the area covered by alveoli (including epithelium and lumen) was measured and the ratio of alveolar area over total organ area was calculated (Figure ), revealing that alveolar area in mutant glands was significantly decreased, on average by 30% (lactation day 3.5) and 20% (lactation day 10.5). Interestingly, the number of alveoli per organ area was not altered (Figure ), suggesting that in mutant glands there is a reduction in the size of alveoli, which could be due to smaller and/or fewer alveolar cells or less milk, even in an active nursing state.
Figure 3 Alterations in alveolar density and secretory activity. (a) Immunohistochemistry against cytokeratin (CK) 18 on sections of wild type (WT) and mutant glands to visualize epithelium (brown). For optimal comparison, a central part of each gland containing (more ...)
To analyze this in more detail, proliferation and apoptosis were investigated in lactating glands. We did not detect any difference in BrdU incorporation between WT and mutant glands (data not shown), nor were shed cells apparent in the lumens (for example, when looking at high magnifications of Figures and ), suggesting no dramatic alterations in cell number. Thus, we examined the glands via electron microscopy to look directly at the secretory activity of alveolar cells. The endoplasmic reticulum forms highly organized, parallel strands, from which secretory vesicles bud to fuse into the alveolar lumen (Figure ). When comparing day 7.5 lactating WT and mutant glands, mutant cells are dominated by parallel regions of thin regular endoplasmic reticulum. In contrast, WT cells contain more dilated reticulum and budding vesicles (arrows), indicating high protein synthesis activity. The result was confirmed in two pairs of day 4.5 lactating mice (not shown). The non-dilated endoplasmic reticulum in c-Myc mutant glands suggests a defect in protein synthesis, at the cellular level.
c-Myc controls biosynthetic activity in the mammary gland
The previous results suggest that c-Myc loss in alveolar cells might cause a general defect in milk production, including milk components and the enzymes involved in their synthesis. This was analyzed in more detail, first by examining mRNA levels of milk proteins and enzymes that are strongly upregulated in lactating mammary glands [43
]. Transcripts encoding: α-lactalbumin (Lalba) and β-casein (Csn2), both milk proteins, the former also the rate-limiting co-factor for lactose synthesis [45
], as well as Δ6 fatty acid desaturase 2 (Fads2), stearoyl-CoA desaturase 2 (Scd2), elongation of very long chain fatty acids (Elovl1), and aldolase C (Aldo3), enzymes involved in lipid synthesis [44
], were measured by semi-quantitative reverse transcription (RT)-PCR. All transcripts were expressed at comparable levels in WT and mutant glands analyzed between lactation day 2.5 and 10.5 (Figure ), including Fads2, Scd2 and Elovl1 that are described Myc targets in other systems (http://www.myc-cancer-gene.org/index.asp
, Zeller et al
]). This suggests that regulation of milk production by c-Myc might occur by a non-transcriptional mechanism.
Figure 4 Altered translation efficiency in mutant mammary glands. (a) Semi-quantitative reverse transcription-polymerase chain reaction (RT-PCR) on α-lactalbumin (Lalba), β-casein (Csn2), Δ6 fatty acid desaturase 2 (Fads2), stearoyl-CoA (more ...)
Next we investigated mRNA translation in WT and mutant glands by performing polysome fractionation on mammary gland lysates obtained at lactation day 4.5. This technique allows the separation of mRNAs along a sucrose gradient depending on their ribosomal load. When overlaying profiles from WT and mutant glands according to their monosome peaks, a change in the average size of polysomes was evident in c-Myc deficient glands, with the peak being shifted to smaller polysomes (Figure , upper panel). Results from one pair of WT and mutant animals are shown; three additional pairs of animals were examined, yielding similar results (data not shown). As a control, we performed polysome fractionations on livers obtained from the females used for generating the mammary gland profiles. WT and mutant mice retain c-Myc in the liver since WAPiCre is not expressed there. The polysome distribution from livers of WT and mutant females was nearly identical (Figure , lower panel), showing that the altered polysome distribution is specific for c-Myc-deficient mammary glands. These results suggest that there is a general reduction in translation efficiency in mammary glands in the absence of c-Myc.
In addition to Pol II targets, c-Myc controls Pol I-mediated rRNA and Pol III-mediated tRNA and 5S rRNA transcription, thereby regulating cellular physiology at multiple levels [1
]. Accordingly, we analyzed a panel of Pol I, II and III c-Myc targets implicated in ribosome biogenesis and translation. The results from qPCR are displayed as relative expression levels in mutant mice, compared with matched WT littermates; the data are from two pairs of mice at the indicated times in lactation (Table ). mRNAs encoding nucleolin and nucleophosmin, which are involved in ribosome biogenesis, mRNAs encoding large and small ribosomal subunit proteins, and the mRNA for poly(A)-binding protein1 (PABPC1), involved in translation, all showed a decrease in samples from mutant females. In particular, the ribosomal protein encoding mRNAs were strongly affected, frequently being more than two-fold downregulated in c-Myc-deficient glands (Table , values below 0.50). Furthermore, the levels of 5S rRNA as well as the rapidly processed 5'-external transcribed spacer of the 45S rRNA precursor [7
], were generally lower in c-Myc mutant glands. This suggests that the decreased translation efficiency in c-Myc mutant glands is due to a general impairment of ribosome biogenesis and translation.
Levels of c-Myc targets involved in ribosome biogenesis and translation
Finally, we examined the translational efficiency, that is, ribosomal load, of specific mRNAs using RNA isolated from each fraction of the polysome gradient. The mRNAs encoding Lalba, Csn2, Fads2, Scd2, Elovl1 and Aldo3 each shifted to smaller polysomes, with the peaks in fractions 7 to 9 in mutant versus 8 to 10 in WT glands (Figure , upper panel, open arrow heads). Interestingly, while each of these transcripts is expressed to the same level in WT and mutant mammary glands (Figure ), this shift clearly shows that they are less efficiently translated. In contrast to the mRNAs encoding proteins directly involved in milk production, the mRNA distribution of β-actin, CK18 and GAPDH along the polysome gradients was essentially the same in WT and mutant glands (Figure , lower panel, open arrow heads). To confirm that the observed reduced translation efficiency results in less protein production in mutant glands, we performed a Western analysis for β-casein on mammary gland lysates (Figure ). Compared with the α-tubulin loading control, there is a clear reduction in casein levels in lysates of mutants compared with WT littermates. Taken together, these results show that a reduction in translation efficiency is likely to be responsible for slower milk production in c-Myc mutant glands.
Delayed proliferative response in c-Myc mutant mammary glands
c-Myc loss has an effect on cell cycle progression and proliferation in many organs [25
]. Thus, we investigated if c-Myc loss affects proliferation during pregnancy. The WAPiCre model is particularly suited for studying proliferation in a second pregnancy since a population of WAPiCre
expressing cells does not undergo a secretory fate, but survives lactation and involution. These cells are termed Pi-MECs (for parity-identified mammary epithelial cells) (see also Smith and Medina [50
]) and function as progenitor cells for epithelium-forming alveolar structures during ensuing rounds of pregnancy and lactation [51
]. In our model, cells in c-mycfl/fl
mice that survive involution will have lost c-Myc due to Cre expression during the first pregnancy. Consistent with these characteristics, the recombined c-myc
allele was detected in non-pregnant, parous females, and in all stages of a second pregnancy (Figure ), in contrast to the first pregnancy where recombination was first detectable at day 14.5 (Figure ). Furthermore, c-myc
mRNA levels are very low during a second pregnancy and lactation in mutant, compared with WT glands (Figure ).
Figure 5 Delayed proliferative response of c-Myc mutant cells in second pregnancy. (a) Detection of recombined allele in mutant glands as described in Figure 1(c). np = non-pregnant, parous. (b) Semi-quantitative reverse transcription-polymerase chain reaction (more ...)
In the normal mammary gland c-myc
mRNA is highest between day 6.5 and day 12.5 of pregnancy then drops to baseline for the remainder of pregnancy and throughout lactation [53
]. In our model, during the first pregnancy Cre activity, hence c-Myc deletion is maximal early in lactation, a time when it has not been possible to detect c-Myc by IHC (data not shown; Klinakis et al
]). However, since the recombined c-myc
allele was detected in all stages of a second pregnancy (Figure ) and c-myc
mRNA levels are very low in mutant glands (Figure ), we performed IHC staining for c-Myc on sections prepared from second pregnancy day 6.5 mammary glands. c-Myc staining was evident in sections prepared from WT females (see Additional file 1
), although not as strong as the day 10.5 embryonic liver positive control [22
]. In contrast, in the mutant glands, c-Myc staining was absent in most of the epithelial clusters. These results clearly show that in c-mycfl/fl
mice c-Myc mRNA and protein are lost.
To monitor proliferation during pregnancy, IHC for Ki-67, which stains all but G0 cells, was performed (Figure , left). Furthermore, cyclin D1, which is preferentially expressed in the mammary gland and is essential for proliferation [54
] was analyzed by IHC (Figure , right). In sections from WT glands the majority of cells were actively cycling at pregnancy day 6.5, displaying positive Ki-67 staining, as well as high levels of cyclin D1. In striking contrast, in mutant glands analyzed on the same day, the majority of cells were Ki-67 negative, and had low or undetectable cyclin D1, showing that most cells were not proliferating. By pregnancy day 14.5, however, the majority of mutant cells were cycling, showing that the slower proliferative response was surmountable. Mutant glands at day 14.5 resembled WT glands at day 6.5, whereas by day 14.5, WT glands displayed advanced development with many lumen-forming, alveolar clusters. Of note, levels of N-myc
were the same as in WT glands, showing that there was no compensation at the mRNA level in glands lacking c-Myc (Figure ). In summary, the results indicate that in the absence of c-Myc, alveolar cells show a delayed proliferative response at the start of pregnancy.
Delayed, but successful differentiation in c-Myc mutant glands
The slower proliferation in the mutant glands resulted in delayed differentiation, which was monitored by IHC against milk proteins (Figure ). While the WT gland at pregnancy day 14.5 was producing milk, as shown by the milk-filled lumen of the alveoli, the small alveolar clusters in the mutant gland were essentially empty and no milk was detected. By pregnancy day 16.5, however, alveoli occasionally contained cytoplasmic lipid droplets (Figure , red circles, insert), indicating that the mutant cells had begun differentiation and milk production. Indeed, despite the delay in development, milk production was successful since mutant mothers were able to nurse their pups after parturition, albeit with pups showing reduced body weight (data not shown). Interestingly, expression of the WAPiCre transgene was also delayed in a second pregnancy (Figure ). A control female (c-Mycfl/+;WAPiCre+) in its second pregnancy and lactation showed scattered Cre staining at pregnancy day 16.5 and ubiquitous staining at lactation day 3.5 (as in Figure ). In contrast, mutant glands displayed almost no Cre-expressing cells at day 16.5 of a second pregnancy and only scattered positive cells at lactation day 3.5. Importantly, at day 10.5 of the second lactation, the mutant glands showed ubiquitous Cre expression, indicating that the transgene had not been silenced (Figure ). These results suggest that in the c-Myc mutant glands the WAPiCre transgene and endogenous milk protein genes show a similar delay in their expression pattern, very likely reflecting the slower proliferative response.
Figure 6 Delayed differentiation of c-Myc mutant glands in a second pregnancy. (a) Immunohistochemistry (IHC) against mouse milk in glands taken from wild type (WT) and mutant mice at second pregnancy day 14.5 and 16.5. Milk proteins are stained in brown, lipid (more ...)
Finally, quantification of the alveolar density showed that during a second round of pregnancy and lactation, c-Myc mutant glands displayed a strongly reduced alveolar area (Figure ). With more than a 40% reduction on lactation day 3.5, this effect is more severe than the 30% decrease observed in a first pregnancy (Figure ). The reduced alveolar area is also evident in whole mount preparations from WT and mutant females obtained at the same time points (Figure ). The results might be explained, in part, by the slower proliferation leading to an incomplete alveolar expansion in the mutant glands (Figure ). In conclusion, the data suggest that c-Myc is dispensable for secretory differentiation, however, due to slower proliferation there is also a delay in differentiation in c-Myc mutant mammary glands.
Effects on progenitor cells in c-Myc mutant glands
Considering the more severe phenotype in the second pregnancy, we performed additional experiments to investigate the role of c-Myc in mammary progenitor cells. A quantification of the alveolar number from three pairs of WT and mutant females, analyzed at lactation day 3.5 of a second pregnancy, revealed a significant reduction in c-Myc mutant glands (Figure ), suggesting that mutant glands start the second pregnancy with fewer alveolar progenitor cells. Importantly, there was no difference in the alveolar number between WT and mutant glands measured at the first lactation (Figure ).
Figure 7 Effects on mammary progenitor cells. (a) Number of alveoli (including mean ± standard deviation) measured in three pairs of animals at second lactation day 3.5, which were used for area measurement in Figure 6(c). Numbers were calculated per analyzed (more ...)
To functionally investigate mammary progenitor cells, we performed reconstitution experiments into cleared mammary fat pads. Pieces of mammary glands from WT and mutant mothers were transplanted into NOD/SCID recipients. Donor glands were taken from lactation day 5.5, a time point when Cre activity is maximal and most cells will have lost c-Myc. Recipients were sacrificed after 8 weeks in order to examine survival and outgrowth potential of mammary progenitor cells. The results from two independent experiments are summarized in Figure . Epithelium from WT donors reconstituted a ductal network in all recipients. A representative outgrowth that filled around 30% of the gland ('+ +') is shown in Figure . In contrast, in 60% of the cases, transplanted epithelium from mutant donors failed to grow out and only rudimentary ductal trees were detected in the recipients (Figure , mutant, left panel). In the cases when mutant epithelium formed ductal outgrowths (Figure , mutant, right panel), these were similar to those formed by WT epithelium. A PCR analysis showed that the recombined allele could be detected in DNA recovered from two positive ('+ +') mutant outgrowths (Figure ), showing that c-Myc-deficient epithelial cells survived and likely contributed to outgrowth formation. In conclusion this suggests that c-Myc has an impact on mammary gland progenitor cell survival and/or proliferation.