Over the first 6 weeks of intestinal development, morphologic parameters increased in a gradual fashion as expected (, panels A–F). Animal weight, small intestinal length, crypt depth and villus height increased in a linear fashion with R2 values of 0.91, 0.99, 0.94 and 0.92, respectively (). In contrast, jejunal circumference and the number of crypts per circumference increased during the first 2 postnatal weeks then plateaued from p21 onward, thus showing a more logarithmic pattern with R2 values of 0.92 and 0.96 ().
In order to explore the hypothesis that developmental crypt fission is driven by increased numbers of ISCs, our first approach was to perform quantitative RT-PCR for several putative mRNA markers of ISC as shown in (panels A–E). Of these, mTert mRNA () surged earliest, at p4, and remained elevated at p7 and p14, then declined to adult levels by p21. Lgr5 and Bmi1 mRNA expression peaked at p14 (), corresponding with the peak of crypt fission. Ascl2 mRNA () was not significantly different from adult at these time points. Additionally, Olfm4 mRNA () remained significantly lower than adulthood at all neonatal time points.
To assess whether there are increased numbers of Lgr5
-expressing ISCs during the period of elevated crypt fission (p7–p21), we utilized the commercially available Lgr5
-EGFP mouse in order to visualize the Lgr5
-expressing cells in the crypts. Representative confocal photomicrographs of Lgr5
-EGFP animals are shown in (panels A–F). At p0, EGFP positive cells were very sporadic in the intervillous region, with only 1–2 EGFP positive cells seen across all observed sections. By p7, a mosaic pattern of EGFP expression was apparent with approximately half the crypts being EGFP positive. In the positive crypts, EGFP localized only to the very base. Over time, as crypt depth increased (), there was a trend toward more EGFP positive cells seen in each crypt (). Starting at p21 (), darker areas in the crypt base correspond with the appearance of Paneth cells and at the adult time point (), fluorescence in the crypt base appeared restricted to several tall columnar cells located throughout the crypt base, consistent with the original report of these mice 
. Quantitative data did not demonstrate more EGFP positive cells per crypt in development compared with adult (). To the contrary there was a gradual trend towards increased numbers of EGFP cells per crypt, from 3.6±0.2 at p7–14 to 4.3±0.2 in adulthood () which parallels the increase in total number of cells per crypt, from 32.7±1.1 to 38.3±0.7 (data not shown). Consequently, the percentage of EGFP positive cells per crypt remained constant () as crypts increased in overall cellularity. The percentage of total crypts that were EGFP positive remained at approximately 50% from p7–p28, then decreased significantly, with only 17.3% ± 1.1 of crypts remaining EGFP positive in adulthood ().
Mature bowel possesses an ability to respond to challenges by augmenting several characteristics after a loss of length or functional capacity 
. This behavior is also evident during development, a time marked by significant increases in both macroscopic and microscopic features of the intestine. A careful analysis of the expansion of the murine intestinal epithelium during postnatal development has been described by Cheng and Bjerknes 
. Although they utilized Swiss albino mice while we studied C57BL/6J animals, overall trends for common measures such as crypt depth and villus height were similar, as was the timing of the peak of crypt fission.
Two parameters included in our study but not reported in the previous study were bowel circumference and the number of crypts per circumference. Bowel circumference has particular clinical relevance as dilation is one of the properties of small intestinal adaptation after massive bowel resection both in mice and in human patients 
. Measuring the crypts per circumference offers additional insight into the rate of crypt formation and the relationship of crypt fission to the postnatal increase of bowel circumference. As can be seen in and , the high rates of crypt fission at p7–21 are temporally associated with the increases in bowel circumference and the number of crypts per circumference. Then as crypt fission declines to low levels at p28 to adulthood, both of these circumference parameters plateau. These observations support the conclusion that during development, just as following resection and damage, crypt fission and the resulting increase in crypt number is imperative for improving functional capacity of the bowel 
The process of crypt fission has been described in rodents and humans 
. The most interesting aspect of this process is its temporal restriction. In each species, there is a peak of fission during development, followed by establishment of a baseline level in adults. In rats, fission peaks on p11 at 10.5% and remains elevated until p25, when it declines to adult baseline levels 
. In humans, a peak is seen between 6–12 months of age at 18% 
. Prior work in Swiss albino mice showed a peak in the number of bifid crypts (35%) at 2–3 weeks of life 
. This corresponds to our current data demonstrating a peak of 22.1% on p14. Interestingly, both in rodents and humans, subsequent decline of crypt fission corresponds with the expected time of weaning (although specific data on weaning were not available from the human study). Certainly, this raises the possibility that factors delivered in the milk may be responsible, at least in part, for the high rates of crypt fission during the suckling period. Crypt fission is also temporally restricted in models of intestinal resection, where a loss of length stimulates ISCs to proliferate and generate additional crypts as seen by increased fission and then a return to homeostatic levels 
. In these examples as in our data, crypt fission appears to reflect ISC expansion and allows for increasing numbers of crypts in a process that is closely regulated during development and after ileocecal resection.
Cheng and Bjerknes have previously quantified the overall proliferative activity of murine intestinal crypts during postnatal development by flow cytometry and found proliferative indices between 7–14% 
. They did not identify changes in proliferation temporally corresponding to the surge in crypt fission. In fact, there was a drop in the percentage of cells in S phase with a nadir on p21, when the rate of fission was still markedly elevated in their mice and significantly elevated in ours.
We identified distinctive patterns of mRNA expression among several putative ISC markers as shown in . The first marker to show elevated expression was mTert
, at p4. This finding was followed by surges in Lgr5
, which peaked at p14, corresponding with the peak of crypt fission. Interestingly, the early increase in mTert
mRNA correlates with recent data indicating that putative ISCs marked by mTert
may give rise to Lgr5+
. In contrast, Olfm4
mRNA levels were significantly lower at all timepoints in development compared with adult while Ascl2
was statistically similar to adulthood at p7–14 during the time of elevated crypt fission and lower than adult levels at the other timepoints (Table S1
). As mentioned, Lgr5
, and Ascl2
are thought to represent the same group of ISCs, marking the crypt base columnar ISC sub-population 
. However, in our study the latter 2 markers were expressed differently from Lgr5
and their expression did not parallel crypt fission levels ( and Table S1
). This observation of divergent mRNA expression among markers of ISCs has been noted before with Olfm4 
. Such divergent behavior stands as a caution that relationships between markers established in adult animals under homeostatic conditions do not necessarily hold during development or during other physiological or pharmacological challenges.
Prior work utilizing SP sorting has demonstrated it to be a population enriched for putative ISCs 
. Hypothesizing that this technique would be similarly useful in the developmental period, we quantitated the SP fraction from p21 mice as compared with adult. The results () showed the% SP at p21 to be 8-fold higher than in adulthood (4.8 vs. 0.58). This correlates closely with crypt fission, which showed a 6-fold difference between p21 and adult (). Although p21 was not the peak of crypt fission, given the fragile nature of neonatal intestinal tissue, we utilized the latest time point at which crypt fission was significantly elevated compared with adult animals in order to obtain the best epithelial preparation possible. This provided additional evidence that there was expansion of ISCs at the time of elevated crypt fission in our model. As cell cycle analysis of hematopoietic SP has shown this fraction to represent cells in G0 [30,31]
, our demonstration of elevated % SP may be linked with the behavior of the quiescent ISC population.
In tissues with complex cellularity such as the intestine, rising levels of mRNAs marking a subset of cells are often taken to indicate an increase in the number of cells producing that particular mRNA. Alternatively, it may reflect an increase in the amount of mRNA produced per cell if the number of cells remains constant. In order to analyze the relationship between ISC mRNA markers and ISC numbers in the developing intestine we would ideally use antibodies to markers of the two putative ISC populations in order to quantitate their numbers. Unfortunately at this time such antibodies are not available. Thus we used the commercially available Lgr5-EGFP mouse.
Taking this approach to assess ISC numbers in the developmental period allowed for visualization of Lgr5
-expressing cells via fluorescence. The increased expression of Lgr5
mRNA () was not seen as an increased percentage of EGFP positive cells per crypt. Interestingly, the percentage of total crypts that were EGFP positive was significantly lower in adulthood compared with all developmental timepoints and highlights the varying degree of mosaic EGFP expression present in this animal. Within the limitations of this particular animal model, the fact that EGFP positive cells were extremely rare at birth may be significant. Given their rarity, we elected not to include that timepoint in statistical analysis. However, with that caveat, there is certainly an obvious increase in EGFP expression at the forming crypt base from p0 to p7 (). This observation is supported by the original paper describing Lgr5
, which reported minimal Lgr5
expression around the time of birth 
hybridization of Lgr5
mRNA in WT mice confirmed the lack of expression at birth (Figure S1
). At 1 week, expression of Lgr5
occurred in all crypts, to support the mosaic nature of the Lgr5
-EGFP mouse, rather than variable expression of Lgr5
itself. Over time, the autoradiographic signal remained strong in the crypts. Although not quantifiable, the number of Lgr5
-expressing cells in the crypts appears to remain relatively constant from p7 into adulthood, which is similar to our EGFP data as shown in . Overall, our data with the Lgr5
-EGFP mouse and Lgr5 in situ
hybridization suggest that Lgr5
-expressing ISC may be involved in the initial formation of crypts during the first postnatal week and then remain relatively constant in number, consistent with their proposed role in steady state turnover of the intestinal epithelium.
Different mouse models offer a variety of circumstances to investigate the process of crypt generation as it relates to ISC biology. Following ileocecal resection, increases of ISC numbers (measured by SP sorting, phospho-β-catenin staining and bromodeoxyuridine label retention) have been reported to precede the increase in crypt fission 
. Interestingly, all of the ISC parameters utilized in that study tend to characterize the “plus 4” quiescent subpopulation of ISCs. To date, specific assessment of the jejunal CBC subpopulation following ICR has not been reported, however other investigators demonstrated no significant increase in ileal Lgr5
expression following 50% small bowel resection 
. In animals treated with doxorubicin, significant surges in crypt fission and% SP are seen with regeneration of crypts starting 1 week post-injury, but there is no increase in either the number of Lgr5
-expressing cells or Lgr5
mRNA levels at similar time points 
Clearly, techniques to directly evaluate actual numbers of ISCs are lacking. With this limitation, surrogates are used to indirectly study ISC expansion. In development, elevated levels of crypt fission correspond with rapid increases in small bowel length, circumference, and the number of crypts per circumference over a discrete time period. We describe patterns of mRNA expression in this time frame for several putative ISC markers and show that Lgr5, Bmi1, and mTert are elevated during and/or preceding the surge in crypt fission while Olfm4 and Ascl2 are not. Crypt fission is a unifying theme in 3 distinct scenarios that involve expansion and/or repair of the intestinal epithelium. Neonatal development is strictly a growth phenomenon while the adaptive process after ileocecal resection is a response of normal bowel to a loss of length elsewhere in the gastrointestinal tract. The response to damage arises from crypts that have sustained an insult whether from radiation or chemotherapeutic agents.
We provide evidence that during development, a heterogeneous population of ISCs, both active (as seen via Lgr5 mRNA expression, Lgr5-EGFP, and Lgr5 in situ hybridization) and quiescent (as seen by mTert mRNA, Bmi1 mRNA and% SP), may play a role in neonatal intestinal growth. While the surge of crypt fission seen at the end of the second postnatal week suggests increased numbers of ISC at this time, we were not able to demonstrate a clear correlation between this surge and either of these populations. This raises the question that both active and quiescent ISC subpopulations may be interacting in a complex manner during neonatal intestinal growth. Further characterization is difficult within the current limitations of the field, namely the lack of any specific, commercially available antibodies for putative ISC markers. All of the measures currently available are surrogates. A better understanding of the regulatory mechanisms controlling crypt fission and the expansion of ISCs has clear translational application to patients suffering from short bowel syndrome and other intestinal pathology. Enhanced techniques to more directly identify, quantify, and manipulate putative ISCs have significant potential. Further characterization of ISC biology may reveal components that are subject to beneficial clinical intervention.