Impaired terminal differentiation of hippocampal granule neurons and defective contextual memory in PC3/Tis21 knockout mice

Abstract

Neurogenesis in the dentate gyrus of the adult hippocampus has been implicated in neural plasticity and memory, but the molecular mechanisms controlling the proliferation and differentiation of newborn neurons and their integration into the synaptic circuitry are still largely unknown. To investigate this issue, we have analyzed the adult hippocampal neurogenesis in a PC3/Tis21-null mouse model. PC3/Tis21 is a transcriptional co-factor endowed with antiproliferative and prodifferentiative properties; indeed, its upregulation in neural progenitors has been shown to induce exit from cell cycle and differentiation. We demonstrate here that the deletion of PC3/Tis21 causes an increased proliferation of progenitor cells in the adult dentate gyrus and an arrest of their terminal differentiation. In fact, in the PC3/Tis21-null hippocampus postmitotic undifferentiated neurons accumulated, while the number of terminally differentiated neurons decreased of 40%. As a result, PC3/Tis21-null mice displayed a deficit of contextual memory. Notably, we observed that PC3/Tis21 can associate to the promoter of Id3, an inhibitor of proneural gene activity, and negatively regulates its expression, indicating that PC3/Tis21 acts upstream of Id3. Our results identify PC3/Tis21 as a gene required in the control of proliferation and terminal differentiation of newborn neurons during adult hippocampal neurogenesis and suggest its involvement in the formation of contextual memories.

Figure 1. Ablation of PC3/Tis21 impairs terminal differentiation of 1- to 5-day-old adult-generated dentate gyrus progenitor cells.

New stem and progenitor cells (A–D) and post-mitotic neurons (E–I) of 1 to 5 days of age, as detected by incorporation of BrdU after five daily injections in P60 PC3/Tis21 ^+/+ and PC3/Tis21 ^−/− mice (referred to as Tis21 throughout the figures), were analyzed by the expression of specific markers through multiple-labeling confocal microscopy. (A) New stem cells (type-1; BrdU/GFAP/nestin-positive) and (B) type-2ab (BrdU/nestin-positive and GFAP-negative) or (C) type-2b (BrdU/nestin/DCX-positive) progenitor cells did not present significant changes of number in PC3/Tis21-null mice. (D) Type-3 progenitor cells (BrdU/DCX-positive nestin-negative) increased significantly, of about 17%. (E) New stage 5 post-mitotic neurons, derived from progenitor cells and identified by NeuN, did not show evident changes at the onset of differentiation (early stage 5, BrdU/DCX/NeuN-positive), while (F) a significant increase of about 2-fold was observed in neurons at a further stage characterized by expression of Calretinin, i.e., in BrdU/Calr/DCX-positive neurons, and (G) in BrdU/Calr/NeuN-positive neurons. (H) Also the whole population of Calretinin-positive neurons of all ages, was significantly increased in PC3/Tis21-null mice (Calr-positive neurons, 25% increase). (I) In contrast, the terminally differentiated stage 6 neurons (BrdU/NeuN-positive and Calr-negative) decreased sharply (65% decrease). (J) Moreover, glial cells, identified as GFAP-positive, did not change in number, indicating that gliogenesis was not affected. The strong increase of stage 5 neurons, in parallel with the strong decrease of stage 6 neurons, suggests that Calretinin-positive neurons accumulate, being unable to attain terminal differentiation at stage 6. The cell numbers in dentate gyrus, shown in (A–J), were measured as described in Materials and Methods and are represented as mean ± SEM of the analysis of three animals per group. ***, p<0.001 vs. PC3/Tis21 ^+/+ dentate gyrus; Student's t test. (K) Representative confocal images showing 1– to 5-day-old stage 5 dentate gyrus neurons positive to BrdU/Calr/NeuN (in orange in the merged image, indicated by arrowheads; single labeling is red, green, blue, respectively), which increase conspicuously in PC3/Tis21-null mice, and of stage 6 new neurons positive to BrdU/NeuN and negative to Calretinin (in purple the merged image, indicated by arrows), which in contrast decrease. Insets show Calretinin-positive and -negative cells at higher magnification. Scale bar, 50 µm; 27 µm in the inset.

Figure 2. Ablation of PC3/Tis21 impairs terminal differentiation of 28-day-old adult-generated dentate gyrus progenitor cells.

(A, B) Neurons of about 28 days of age, as defined by BrdU birthdating, expressing either (A) the immature marker DCX (BrdU/DCX/NeuN-positive), or (B) Calretinin - that labels neurons not yet terminally differentiated - (BrdU/Calr/NeuN-positive), were significantly increased in PC3/Tis21-null mice (50% and 2.2 fold increase, respectively). (C, D) In contrast, terminally differentiated 28 day-old neurons identified either as (C) BrdU/NeuN-positive and Calr-negative or as (D) BrdU/Calbindin/NeuN-positive cells, decreased sharply (about 38% and 35%, respectively). (E) The population of NeuN-positive neurons, analyzed as a whole, also decreased significantly, albeit to a lower extent (27%). Thus, the impairment of terminal differentiation observed in 1- to 5-day-old neurons (Figure 1) was still present also in older neurons of 4 weeks. Cell numbers are represented as mean ± SEM of the analysis of three animals per group. *, p<0.05 or **, p<0.01 or ***, p<0.001 vs. PC3/Tis21 ^+/+ dentate gyrus; Student's t test. (F) Representative confocal images showing 4-week-old stage 5 dentate gyrus neurons positive to BrdU/Calretinin/NeuN, which increase in PC3/Tis21-null mice (in green or pink in the merged image, indicated by arrowheads; single labeling is red, green, blue, respectively), and of stage 6 new neurons positive to BrdU/NeuN and negative to Calretinin (in purple, indicated by arrows), whose number decreases. It is also detectable an increase of the whole population of Calretinin-positive cells in PC3/Tis21-null mice (compare the right panels). Insets show single Calretinin-positive or -negative cells. Scale bar, 45 µm; 15 µm in the inset.

Figure 3. Dentate gyrus progenitor cells proliferate more in mice ablated of PC3/Tis21.

(A) In P60 PC3/Tis21-null mice the total number of progenitor cells in the dentate gyrus entering in S-phase, identified by a short pulse of BrdU, increased significantly, of about 30%. (B) A similar, significant increase occurred also for the total number of dividing progenitors in the dentate gyrus, measured as Ki67-positive cells, and (C) for the ratio between cells in S-phase (BrdU⁺) and all dividing progenitors (Ki67⁺). This ratio, with a net increase of progenitor cells entering in S-phase with respect to the number of dividing cells, indicates that the length of cell cycle (G1 phase) was shorter in PC3/Tis21-null mice. (D) Consistently, no difference was observed in the ratio between cells in mitosis (PH3⁺) and all dividing progenitors (Ki67⁺). An analysis of the types of dividing progenitor cells types indicates that while the number of (E) type-1 (Ki67/GFAP/ nestin-positive) and (F) type-2ab (Ki67/nestin-positive and GFAP-negative) did not change significantly, the subpopulation (G) type-2b (Ki67/nestin/DCX-positive) and also (H) type-3 (Ki67/DCX-positive and nestin-negative) increased, in the latter case significantly, of about 50% and 80%, respectively. These results are consistent with a model where PC3/Tis21 physiologically inhibits proliferation and favors differentiation. Cell numbers are represented as mean ± SEM of the analysis of three animals per group. **, p<0.01 or ***, p<0.001 vs. PC3/Tis21 ^+/+ dentate gyrus; Student's t test. (I) Representative confocal images showing dentate gyrus progenitors either entering in S-phase or dividing, labeled by a 2 hours BrdU pulse (in red), or by Ki67 (in green), respectively. In both cases, their number increases in PC3/Tis21-null mice, as compared to WT mice. Scale bar, 50 µm. (J) A progenitor cell positive to phospho-histone H3, marker of the G2/M cell cycle phase, in dentate gyrus of PC3/Tis21-null mice. PH3, Ki67 and nuclei (labeled by Hoechst 33258) are in red, green and blue, respectively. SGZ, subgranular zone. Scale bar, 30 µm.

Figure 4. Increase of proliferating progenitor cells and decrease of terminally differentiated granule neurons in the immature dentate gyrus of PC3/Tis21-null mice at P14.

(A) The whole population of progenitor cells entering in S-phase, labeled by a short pulse of BrdU (1hour) and by the progenitor cell marker Sox2 (BrdU/Sox2-positive), increased significantly in PC3/Tis21-null mice (1.6-fold). (B, C) The same significant increase is observed for the number of dividing progenitor cells, identified either as (B) Ki67-positive, or as (C) Ki67/Sox2-positive cells (1.5-fold). (D) The newly generated neurons (1 hour pulse of BrdU) decreased sharply (20%). Thus, also in the immature PC3/Tis21-null dentate gyrus prevails a condition of increased proliferation of progenitor cells and reduced differentiation of new neurons. Cell numbers are represented as mean ± SEM of the analysis of three animals per group. *, p<0.05 or **, p<0.01 vs. PC3/Tis21 ^+/+ dentate gyrus; Student's t test. (E) Representative confocal images of the dentate gyrus in P14 mice, showing cells double positive to BrdU/Sox2, in lower number in PC3/Tis21-null mice compared to WT. The merged images show that the new neurons incorporating BrdU (in red-pink) are localized quite exclusively in the proliferative subgranular zone and are virtually all positive also to the marker of proliferating progenitor cells Sox2. Scale bar, 75 µm.

Figure 5. Cell survival analysis in PC3/Tis21 ^−/− and PC3/Tis21 ^+/+ mice.

(A–D) Cell survival was analyzed in the whole cell population of dentate gyrus of P60 normal and mutant mice and also in the different subpopulations of progenitor cells and neurons, using Caspase-3 as apoptotic marker. (A) A significant increase of the number of cells undergoing apoptosis was observed in all cells of dentate gyrus (1.6-fold increase) of PC3/Tis21-null mice, and more specifically in (B) type-2b and type-3 progenitor cells, measured as percentage of Caspase-3-DCX-positive and NeuN-negative cells in the population of DCX-positive and NeuN-negative cells (3.2-fold increase). (C) The percentage of undifferentiated apoptotic stage 5 neurons (Caspase-3/Calretinin/NeuN-positive in Calretinin/NeuN-positive) presented a not significant increase, while (D) a significant reduction was evident in the percentage of apoptotic stage 6 neurons in PC3/Tis21-null mice (Caspase-3/Calretinin-negative NeuN-positive in Calretinin-negative NeuN-positive; about 54% decrease). Thus, apoptosis was more frequent in progenitor cells type-2b and type-3 whose proliferation was increased. Cell numbers are represented as mean ± SEM of the analysis of three animals per group. *, p<0.05, or ***, p<0.001 vs. PC3/Tis21 ^+/+ dentate gyrus; Student's t test. (E) Confocal microscopy fields showing examples of apoptotic cells in the dentate gyrus of P60 mice, indicated by arrows, i.e., a type-2b/3 progenitor cell (Caspase-3/DCX-positive and/NeuN-negative) in PC3/Tis21-null mice, and a stage 6 neuron (Caspase-3/NeuN-positive and Calretinin-negative) in PC3/Tis21 WT mice. Scale bar, 45 µm.

Figure 6. PC3/Tis21 ^−/− mice show selective impairment of contextual memory and preferential recruitment of stage 5 neurons into dentate gyrus after both spatial and contextual behavioral task.

(A) Escape latency (sec) throughout the Morris water maze behavioral test, to reach the hidden platform during the 6-day acquisition phase. (B) Time (sec) spent in the quadrants during the probe trial, when the hidden platform, located in the target quadrant (TG) during the acquisition phase, was removed. Both PC3/Tis21 ^−/− and PC3/Tis21 ^+/+ mice spent a significantly greater amount of time in the TG quadrant, compared to the other quadrants. AR, adiacent right; AL, adiacent left; OP, opposite. (C) Contextual fear conditioning test: percentage of time spent in freezing behavior by PC3/Tis21 ^−/− and PC3/Tis21 ^+/+ during both training (left) and test (right). During the training, no significant differences in the level of freezing between groups were observed both before (Pre-US) and after (Post-US) foot-shock administration. During the test, performed 24 h after training, PC3/Tis21 ^−/− mice showed a significant reduction of freezing behavior compared to control mice. *, p<0.05, ANOVA. (D) Inhibitory avoidance test: mean latency to step-through into the dark compartment by PC3/Tis21 ^−/− and PC3/Tis21 ^+/+ mice during both training (left) and test (right). During training, no significant differences were observed in the step-through latency between groups. During the test, performed 24 h after training, PC3/Tis21 ^−/− mice showed a significant reduction of the step-through latency compared to that of control mice. *, p<0.05, ANOVA. (E) The activation into memory circuits of neurons within the whole dentate gyrus neuronal population (i.e., comprising stage 5 and 6) after the Morris water maze, measured as number of c-fos/NeuN-positive cells, was similar in PC3/Tis21-null and WT mice, while (F) after the contextual fear conditioning was significantly lower in PC3/Tis21-null mice. (G) The activation into memory circuits of stage 5 neurons, measured as number of c-fos/Calretinin/NeuN-positive cells, occurred quite exclusively in PC3/Tis21-null mice after the Morris water maze test and (H) also after the contextual fear conditioning test (being no c-fos/Calretinin/NeuN-positive neuron detectable in WT mice after the Morris water maze, or only few after the contextual fear conditioning test). This suggests a preferential recruitment in spatial memory networks of stage 5 neurons of PC3/Tis21-null mice. Analyses shown in (E–H) were performed in PC3/Tis21 ^−/− and WT mice 1.5 hours after the end of behavioral test, expected to activate responsive neurons inducing c-fos expression. Cell numbers are represented as mean ± SEM of the analysis of three animals per group. **, p<0.01 or ***, p<0.001 vs. PC3/Tis21 ^+/+ dentate gyrus; Student's t test. (I) Representative confocal images showing stage 5 dentate gyrus neurons expressing c-fos, following behavioral training by the Morris water maze test, which are detectable only in PC3/Tis21-null mice (c-fos/Calretinin/NeuN-positive cells, in orange in the merged image, are indicated by arrowheads; single labeling is blue, red, green, respectively), or Calretinin/NeuN-positive c-fos-negative cells (indicated by arrows). Scale bar, 75 µm.

Figure 7. Binding of PC3/Tis21 to the Id3 promoter and corresponding decrease of Id3 mRNA.

(A–D) A PC12 cell line carrying an inducible PC3 (rat) cDNA under control of the tet-off system (i.e., a tetracycline responsive element TRE-PC3, activated by a tetracycline-regulated transactivator, CMV-tTA, in the absence of doxycycline) was used to analyze the binding of PC3 to the Id3 promoter, by chromatin immunoprecipitation (ChIP). (A) PC12 cells without (white columns) or with expression of exogenous PC3 (activated a week before by withdrawal of 2 µg/ml of doxycycline; grey columns) were exposed to NGF (100 ng/ml) for 1 and 48 hours. The levels of exogenous PC3 increased considerably in the absence of NGF (t 0), and even more as the treatment with NGF proceeded. Endogenous PC3 was also transiently i51duced by NGF after 1 hour, as expected (Bradbury et al., 1991). (B) ChIP analysis of PC3 binding to the Id3 promoter in PC12 cells, treated as shown in (A). A scheme above the graph describes the Id3 gene and the promoter region analyzed, 700 nt 5′ to the transcription start. The amount of Id3 promoter present in immunoprecipitates obtained using anti-PC3 antibody is quantified by real time PCR and is expressed as fold enrichment (ratio of the percentage of the PC3-immunoprecipitated amount of Id3 promoter detected in the input cell lysates to the percentage of the normal serum-immunoprecipitated amount detected in the input cell lysates). The binding of exogenous PC3 (grey columns) to the Id3 promoter increased in correlation with the increase of exogenous PC3 mRNA levels (panel A). (C) No binding above background by PC3 was observed on the negative control muscle creatin kinase (MCK) promoter. (D) Id3 mRNA levels in differentiating PC12 cells (1 and 48 hours after NGF) decreased significantly, in correlation with the increase of binding of endogenous PC3 to the Id3 promoter. Mean ± SEM fold increases (mRNA) or fold enrichments (ChIP) are from three independent experiments performed in parallel for ChIP and mRNA analysis, using duplicate cultures of the same cells. *, p<0.05, or ** p<0.001 vs. the corresponding time point of PC12 cells without exogenous PC3, Student's t test. (E) Id3 promoter activity in differentiating PC12 cells (1 and 48 hours after NGF) decreased significantly in correlation with the increase of binding of endogenous PC3 to the Id3 promoter. The Id3 promoter construct comprised 1592 nt 5′ to the putative transcription start, placed upstream of a luciferase reporter (construct pGL3-Id3-prom/-1592). Luciferase activity is represented as mean ± SEM fold increase from four experiments. *, p<0.05 vs. the corresponding time point of PC12 cells without exogenous PC3, Student's t test. (F) The expression of Id3 mRNA in dentate gyrus of P14 mice, as detected by in situ hybridization, increased considerably in PC3/Tis21-null mice, in the subgranular zone (SGZ) as well as in the granule cell layer (GCL). A representative image from one of three independent in situ experiments is shown. Scale bar, 80 µm. (G) The Id3 mRNA levels in the dentate gyrus of P14 mice, measured by real time PCR after laser capture microdissection of the area, increased significantly in PC3/Tis21-null mice. Data are shown as mean ± SEM fold increase with respect to values obtained in WT, from three independent experiments. **, p<0.01 vs. WT mice, Student's t test (performed on data normalized to the endogenous controls but not yet relativized as fold-expression).

Figure 8. Increase of Id3-positive proliferating progenitor cells and decrease of Id3-positive differentiated neurons in the dentate gyrus of P14 PC3/Tis21-null mice.

(A) In PC3/Tis21-null mice the total population of Id3-positive cells in the dentate gyrus and (B) the population of Id3-positive progenitor cells entering in S-phase, labeled by BrdU (1hour pulse) and by Sox2 (BrdU/Sox2/Id3-positive cells), increased significantly (1.31-fold and 1.56-fold, respectively). (C) Also the percentage of Id3-positive cells among the new proliferating progenitor cells (BrdU/Sox2/Id3-positive in BrdU/Sox2-positive) presented a significant increase. (D) In contrast, the new differentiating neurons expressing Id3 (BrdU/NeuN/Id3-positive neurons) decreased considerably, and (E) a significant decrease was observed also for the total population of terminally differentiated neurons co-expressing Id3 and Calbindin. Cell numbers are represented as mean ± SEM of the analysis of three animals per group. *, p<0.05 or ***, p<0.001 vs. PC3/Tis21 ^+/+ dentate gyrus; Student's t test. (F, G) Representative confocal images of the dentate gyrus in P14 mice; (F) The new proliferating progenitor cells expressing Id3, identified also by BrdU and Sox2 (BrdU/Sox2/Id3-positive, indicated by arrowheads; single labeling is red, blue, green, respectively) increase significantly in PC3/Tis21-null mice. It is also evident that the total number of Id3-positive cells increases greatly, with a localization expanded to the whole area of the granule cell layer, up to its outer third (delimited by the white broken line). Scale bar, 45 µm. (G) The whole population of terminally differentiated neurons, identified by Calbindin and co-expressing Id3, in PC3/Tis21-null mice is restricted toward the outer border of the dentate gyrus, in correspondence with the expansion to that area of the Id3-expressing cells. The white broken line delimits the inner and outer borders of dentate gyrus. Scale bar, 45 µm. (H) A model proposed for PC3/Tis21 activity during neuronal differentiation (see Discussion). Delta1 is represented upstream of PC3/Tis21, whereas cyclin D1 downstream, in agreement with previous findings , .