Dual role for Headcase in hemocyte progenitor fate determination in Drosophila melanogaster

Abstract

The hematopoietic organ of the Drosophila larva, the lymph gland, is a simplified representation of mammalian hematopoietic compartments, with the presence of hemocyte progenitors in the medullary zone (MZ), differentiated hemocytes in the cortical zone (CZ), and a hematopoietic niche called the posterior signaling centre (PSC) that orchestrates progenitor differentiation. Our previous work has demonstrated that the imaginal cell factor Headcase (Hdc, Heca) is required in the hematopoietic niche to control the differentiation of hemocyte progenitors. However, the downstream mechanisms of Hdc-mediated hematopoietic control remained unknown. Here we show that Hdc exerts this function by negatively regulating the insulin/mTOR signaling in the niche. When Hdc is depleted in the PSC, the overactivation of this pathway triggers reactive oxygen species (ROS) accumulation and, in turn, the differentiation of effector lamellocytes non-cell-autonomously. Although overactivation of insulin/mTOR signaling normally leads to an increase in the size of the hematopoietic niche, this effect is concealed by cell death caused by hdc loss-of-function. Moreover, we describe here that hdc silencing in progenitors causes cell-autonomous ROS elevation and JNK pathway activation, resulting in decreased MZ size and differentiation of lamellocytes. Similarly to the PSC niche, knocking down hdc in the MZ also leads to caspase activation. Notably, depleting Hdc in the progenitors triggers proliferation, an opposing effect to what is observed in the niche. These findings further our understanding of how progenitor maintenance in the larval lymph gland is controlled autonomously and non-cell-autonomously, and point towards new mechanisms potentially regulating HSC maintenance across vertebrates.

Fig 1. Hdc negatively regulates the insulin/mTOR pathway in the PSC.

(A-F) Lamellocytes (red) are absent from control lymph glands (Pcol85-Gal4/+) (n = 14) (A), while they are present when hdc is silenced in the PSC (Pcol85-Gal4,UAS-hdcRNAi/+) (n = 14) (B), its partner unk is silenced (Pcol85-Gal4/UAS-unkRNAi) (n = 12) (C), hdc and unk are silenced together (Pcol85-Gal4,UAS-hdcRNAi/UAS-unkRNAi) (n = 18) (D), a constitutively active Pi3K is expressed (UAS-Pi3K92E.CAAX/+; Pcol85-Gal4/+) (n = 16) (E), or the negative insulin pathway regulator Pten is silenced (Pcol85-Gal4/+; UAS-PtenRNAi/+) (n = 16) (F). (G-H) Lamellocyte differentiation in lymph glands of col>hdcRNAi larvae is rescued when simultaneously either Akt (Pcol85-Gal4,UAS-hdcRNAi/+; UAS-AktRNAi/+) (n = 22) (G) or raptor (Pcol85-Gal4,UAS-hdcRNAi/+; UAS-raptorRNAi/+) (n = 14) is silenced (H). (I) Overexpression of hdc rescues lamellocyte differentiation in col>PtenRNAi larvae (Pcol85-Gal4/+; UAS-PtenRNAi/UAS-hdc.S) (n = 16). n refers to the number of lymph gland lobes analyzed. Nuclei are visualized by DAPI (blue). Scale bar: 20 μm. (J) Scatter plot of the number of lamellocytes per lymph gland lobe in the genotypes presented in panels (A-I). Each dot in the graph represents one lymph gland lobe. Data were analyzed using ANOVA with Tukey’s test for multiple comparisons, ** p ≤ 0.01, *** p ≤ 0.001, **** p ≤ 0.0001. (K-L’) pAkt antibody staining (red) detects higher levels of pAkt in the niche (col>GFP positive cells, green) when hdc is silenced (Pcol85-Gal4,UAS-2xEGFP/UAS-hdcRNAi) (n = 16) (L-L’) in comparison to control (Pcol85-Gal4,UAS-GFP/+) (n = 14) (K-K’). n refers to the number of lymph gland lobes analyzed. Scale bar: 20 μm. (M) A scatter dot plot showing the fold change (average = 4.5 folds) increase in the mean fluorescence intensity (MFI) of pAkt in the PSC (col>GFP positive cells) of col>hdcRNAi (Pcol85-Gal4,UAS-2xEGFP/UAS-hdcRNAi) larvae in comparison to the control (Pcol85-Gal4,UAS-2xEGFP/+). Each dot in the graph represents a PSC from one lobe. Data were analyzed using two-tailed unpaired Student’s t-test, **** p ≤ 0.0001.

Fig 2. hdc silencing causes cell death and cell cycle arrest in the PSC.

(A-C) An enlargement of the PSC (number of col>GFP positive cells) is observed when the insulin/mTOR pathway is activated by expressing Pi3KCa (UAS-Pi3K92E.CAAX; Pcol85-Gal4,UAS-2xEGFP/+) (average number of PSC cells = 75, n = 16) (B) or silencing Pten (Pcol85-Gal4,UAS-2xEGFP/+;UAS-PtenRNAi) (average number of PSC cells = 66, n = 22) (C) in comparison to the control (Pcol85-Gal4,UAS-2xEGFP/+) (average number of PSC cells = 31, n = 14) (A). (D-F) Silencing hdc alone does not alter PSC size (Pcol85-Gal4,hdcRNAi/+; UAS-2xEGFP/+) (average number of PSC cells = 28, n = 14) (D), while simultaneous overexpression of the apoptosis inhibitor p35 increases PSC cell numbers (Pcol85-Gal4,hdcRNAi/+; UAS-2xEGFP/UAS-p35) (average number of PSC cells = 59, n = 16) (E), a phenotype not observed in case p35 is overexpressed alone (Pcol85-Gal4,UAS-2xEGFP/+; UAS-p35/+) (average number of PSC cells = 36, n = 14) (F). (G) Silencing hdc simultaneously with Pten reduces PSC size of col>PtenRNAi larvae to normal (Pcol85-Gal4,hdcRNAi/+; UAS-2xEGFP/UAS-PtenRNAi) (average number of PSC cells = 30, n = 14) (blue: nuclei, green: PSC, red: lamellocytes). n refers to the number of lymph gland lobes analyzed. Scale bar: 20 μm. (H) A scatter dot plot showing PSC cell number in larvae from the genotypes presented in panels (A-G). Each dot in the graph represents a PSC from one lymph gland lobe. Data were analyzed using ANOVA with Tukey’s test for multiple comparisons, ** p ≤ 0.01, **** p ≤ 0.0001, ns: non-significant. (I-J) Apoptotic cells (Dcp1 positive, red) can be observed at a higher frequency when hdc is silenced in the niche (Pcol85-Gal4,UAS-2xEGFP/UAS-hdcRNAi) (n = 16) (J) in comparison to the control (Pcol85-Gal4,UAS-2xEGFP/+) (n = 14) (I). n refers to the number of lymph gland lobes analyzed. Scale bar: 20 μm. (K) A scatter dot plot quantifying the number of Dcp1 positive cells in the niches (col>GFP positive cells) of genotypes presented in the panels (I-J). Each dot in the graph represents a PSC from one lymph gland lobe. Data were analyzed using two-tailed unpaired Student’s t-test, ** p ≤ 0.01. (L-M) FUCCI cell cycle reporter pattern (green: G1 phase, red: S phase, yellow: G2/M phase) in the niche of col>hdcRNAi (Pcol85-Gal4,UAS-hdcRNAi/+; UAS-EGFP::E2F1^1-230,UAS-mRFP1::CycB^1-266) (n = 20) (M) in comparison to the control (Pcol85-Gal4/+; UAS-EGFP::E2F1^1-230,UAS-mRFP1::CycB^1-266) (n = 22) (L). n refers to the number of lymph gland lobes analyzed. Scale bar: 20 μm. (N) A bar graph showing the mean and standard deviation of the number of niche cells in the G2/M phase in the genotypes presented in the panels (L-M). Data were analyzed using two-tailed unpaired Student’s t-test, * p ≤ 0.05. (O-P) Dividing cells (pH3 positive, red) can be observed at a lower frequency when hdc is silenced in the niche (Pcol85-Gal4,UAS-2xEGFP /UAS-hdcRNAi) (n = 20) (P) in comparison to the control (Pcol85-Gal4,UAS-2xEGFP/+) (n = 24) (O). n refers to the number of lymph gland lobes analyzed. Scale bar: 20 μm. (Q) A scatter dot plot quantifying the number of pH3 positive cells in the niches (col>GFP positive cells) of genotypes presented in the panels (O-P). Each dot in the graph represents a PSC from one lymph gland lobe. Data were analyzed using two-tailed unpaired Student’s t-test, ** p ≤ 0.01.

Fig 3. hdc silencing results in cellular stress in the hematopoietic niche.

(A-B’) Silencing hdc results in the activation of the gstD-GFP reporter in the PSC (Col antibody positive area) (Pcol85-Gal4,UAS-hdcRNAi/gstD-GFP) (n = 14) (B-B’), in comparison to the control (Pcol85-Gal4/gstD-GFP) (n = 20) (A-A’) (blue: nuclei, green: ROS, magenta: Collier). n refers to the number of lymph gland lobes analyzed. Scale bar: 20 μm. (C) A scatter dot plot showing the fold change (average = 3.2 folds) increase in the mean fluorescence intensity (MFI) of gstD-GFP in the PSC of col>hdcRNAi (Pcol85-Gal4,UAS-hdcRNAi/gstD-GFP) larvae in comparison to the control (Pcol85-Gal4/gstD-GFP). Each dot in the graph represents a PSC from one lobe. Data were analyzed using two-tailed unpaired Student’s t-test, **** p ≤ 0.0001. (D-E’) Silencing hdc induces the transcription of Thor in the PSC (col>GFP positive area) as detected by an anti-lacZ staining for the Thor-lacZ reporter (Pcol85-Gal4,UAS-hdcRNAi/Thor-lacZ) (n = 10) (E-E’) in comparison to the control (Pcol85-Gal4/Thor-lacZ) (n = 10) (D-D’) (blue: nuclei, green: PSC, red: Thor-LacZ). n refers to the number of lymph gland lobes analyzed. Scale bar: 20 μm. (F) A scatter dot plot showing the fold change (average = 3.8 folds) increase in MFI of Thor-LacZ in the PSC cells of col>hdcRNAi (Pcol85-Gal4,UAS-hdcRNAi/Thor-lacZ) larvae compared to the control (Pcol85-Gal4/Thor-lacZ). Each dot in the graph represents a PSC from one lobe. Data were analyzed using two-tailed unpaired Student’s t-test, **** p ≤ 0.0001. (G-I) Overexpression of Cat (Pcol85-Gal4,UAS-hdcRNAi/UAS-Cat) (n = 18) (H) and foxo (Pcol85-Gal4,UAS-hdcRNAi/UAS-foxo) (n = 20) (I) rescues lamellocyte differentiation in the lymph glands of col>hdcRNAi (Pcol85-Gal4,UAS-hdcRNAi/+) larvae (n = 14) (G). (J-K) Silencing foxo (Pcol85-Gal4,UAS-hdcRNAi/+; UAS-foxoRNAi/+) (n = 20) enhances lamellocyte differentiation in col>hdcRNAi lymph glands (K), while silencing foxo alone (Pcol85-Gal4/+; UAS-foxoRNAi/+) does not lead to lamellocyte differentiation (n = 18) (J). (L) Silencing spi does not affect lamellocyte differentiation in col>hdcRNAi lymph glands (Pcol85-Gal4,UAS-hdcRNAi/UAS-spiRNAi) (n = 14) (blue: nuclei, red: lamellocytes). n refers to the number of lymph gland lobes analyzed. Scale bar: 20 μm. (M) A scatter dot plot showing the number of lamellocytes per lymph gland lobe in the genotypes presented in panels (G-L). Each dot in the graph represents one lymph gland lobe. Data were analyzed using ANOVA with Tukey’s test for multiple comparisons, *** p ≤ 0.001, **** p ≤ 0.0001, ns: non-significant.

Fig 4. A cell-autonomous role of Hdc in the medullary zone.

(A-A”) hdc>GFP shows complementary expression to the CZ marker Hml:DsRed (UAS-mCD8::GFP; hdc¹⁹-Gal4/Hml:DsRed) (n = 16) (blue: nuclei, green: hdc, red: CZ). Scale bar: 20 μm. (B-C”) hdc expression decreases drastically in the lymph gland 16 hpi (n = 14) (C-C”) in comparison to naive control lymph glands (UAS-mCD8::GFP/+; hdc¹⁹-Gal4/+) (n = 10) (B-B”) (blue: nuclei, green: hdc, red: Col). Scale bar: 20 μm. (D-E) A scatter dot plot showing the fold change decrease in MFI of hdc>GFP in per lymph gland lobe (average = -12.5 folds) (D) and per PSC (Col antibody positive area) (average = -6.25 folds) (E) of wasp infested (16 hpi) larvae compared to the control (UAS-mCD8::GFP/+; hdc¹⁹-Gal4/+). Each dot in the graph represents one anterior lobe. Data were analyzed using two-tailed unpaired Student’s t-test, *** p ≤ 0.001. (F-G) Silencing hdc in the MZ of the lymph gland leads to lamellocyte differentiation (UAS-hdcRNAi/+; domeMESO-GAL4,UAS-2xEGFP/+) (n = 20) (G), while lamellocytes are normally not detected in the control (domeMESO-GAL4,UAS-2xEGFP/+) (n = 20) (F) (blue: nuclei, green: MZ, red: lamellocytes). (H-I) Similarly to the control (Tep4-Gal4; UAS-2xEGFP/+) (n = 14) (H), lamellocytes are not observed when hdc is silenced using the Tep4-Gal4 specific for core progenitors (Tep4-Gal4; UAS-hdcRNAi/+; UAS-2xEGFP/+) (n = 14) (I) (blue: nuclei, green: core progenitors, red: lamellocytes). (J-K) Like in the control (CHIZ-Gal4/+) (n = 12) (J), lamellocytes do not differentiate in the lymph gland when hdc is silenced using the CHIZ-Gal4 specific for the IZ (CHIZ-Gal4/UAS-hdcRNAi) (n = 16) (K) (blue: nuclei, red: lamellocytes). n refers to the number of lymph gland lobes analyzed. Scale bar: 20 μm. (L) A scatter dot plot showing the number of lamellocytes per lymph gland lobe in the genotypes presented in panels (F-K). Each dot in the graph represents one lymph gland lobe. Data were analyzed using ANOVA with Tukey’s test for multiple comparisons, **** p ≤ 0.0001, ns: non-significant.

Fig 5. Hdc loss in the MZ causes lamellocyte differentiation through cell-autonomous mechanisms.

(A-B) Silencing hdc in the MZ leads to higher levels of ROS in the anterior lobes but not in the PSC as visualized by the gstD-GFP reporter (UAS-hdcRNAi/gstD-GFP; domeMESO-GAL4/+) (n = 16) (B) in comparison to the control (gstD-GFP/+; domeMESO-GAL4/+) (n = 12) (A) (green: ROS). n refers to the number of lymph glands analyzed. Scale bar: 20 μm. (C-D) Scatter dot plots showing the fold change increase in the MFI of gstD-GFP in the anterior lymph gland lobes (average = 1.8 folds) (C) and the indifference in MFI of gstD-GFP in the PSC of domeMESO>hdcRNAi (UAS-hdcRNAi/gstD-GFP; domeMESO-GAL4/+) larvae in comparison to the control (gstD-GFP/+; domeMESO-GAL4/+) (D). Each dot in the graphs represents one anterior lobe. Data were analyzed using two-tailed unpaired Student’s t-test, **** p ≤ 0.0001, ns: non-significant. (E-I) Overexpression of Cat (UAS-hdcRNAi/UAS-Cat; domeMESO-GAL4,UAS-2xEGFP/+) (n = 16) (E) or foxo (UAS-hdcRNAi/UAS-foxo; domeMESO-GAL4,UAS-2xEGFP/+) (n = 18) (F) or expression of a dominant negative version of bsk (UAS-hdcRNAi/+; domeMESO-GAL4,UAS-2xEGFP/UAS-bsk53R) (n = 14) (G), or overexpression of E-cad (UAS-hdcRNAi/+; domeMESO-GAL4,UAS-2xEGFP/UAS-E-cad) (n = 18) (H) or expression of a dominant negative version of EGFR (UAS-hdcRNAi/UAS-EGFR.DN; domeMESO-GAL4,UAS-2xEGFP/UAS-EGFR.DN) (n = 18) (I) rescues lamellocyte differentiation in hdc silenced background (blue: nuclei, green: MZ, red: lamellocytes). n refers to the number of lymph glands analyzed. Scale bar: 20 μm. (J) The number of lamellocytes per lymph gland lobe from the genotypes presented in panels (E-I) in addition to the control (domeMESO-GAL4,UAS-2xEGFP/+) (n = 20) and domeMESO>hdcRNAi (UAS-hdcRNAi/+; domeMESO-GAL4,UAS-2xEGFP/+) (n = 20). n refers to the number of lymph glands analyzed. Each dot in the graph represents one lymph gland lobe. Data were analyzed using ANOVA with Tukey’s test for multiple comparisons, **** p ≤ 0.0001, ns: non-significant. (K) The percentage of domeMESO positive area per anterior lymph gland lobe from the genotypes presented in panels (E-I) in addition to the control (domeMESO-GAL4,UAS-2xEGFP/+) (n = 20) and domeMESO>hdcRNAi (UAS-hdcRNAi/+; domeMESO-GAL4,UAS-2xEGFP/+) (n = 20). n refers to the number of lymph glands analyzed. Each dot in the graph represents one lymph gland lobe. Data were analyzed using ANOVA with Tukey’s test for multiple comparisons, * p ≤ 0.05, *** p ≤ 0.001, **** p ≤ 0.0001. (L-M) Silencing hdc in the MZ increases the number of puc-lacZ positive cells (UAS-hdcRNAi/+; domeMESO-Gal4,UAS-2xEGFP/puc-lacZ) (n = 16) (M) in comparison to the control (domeMESO-Gal4,UAS-2xEGFP/puc-lacZ) (n = 30) (L) (red: puc-LacZ). Scale bar: 20 μm. (N) A scatter dot plot showing fold change increase in puc-LacZ cell number (average = 4.4 folds) per anterior lobe of domeMESO>hdcRNAi (UAS-hdcRNAi/+; domeMESO-Gal4,UAS-2xEGFP/puc-lacZ) larvae compared to the control (domeMESO-Gal4,UAS-2xEGFP/puc-lacZ). Each dot in the graph represents one anterior lobe. Data were analyzed using two-tailed unpaired Student’s t-test, ns: non-significant. (O-P) Silencing hdc in the MZ increases the number of Dcp1 positive cells (UAS-hdcRNAi/+; domeMESO-Gal4,UAS-2xEGFP/+) (n = 12) (P) in comparison to the control (domeMESO-Gal4,UAS-2xEGFP/+) (n = 22) (O) (red: Dcp1). Scale bar: 20 μm. (Q) A scatter dot plot showing fold change increase in the number of Dcp1 positive cells (average = 9 folds) per anterior lobe of domeMESO>hdcRNAi (UAS-hdcRNAi/+; domeMESO-Gal4,UAS-2xEGFP/+) larvae compared to the control (domeMESO-Gal4,UAS-2xEGFP/+). Each dot in the graph represents one anterior lobe. Data were analyzed using two-tailed unpaired Student’s t-test, **** p ≤ 0.0001. (R) A scatter dot plot showing the insignificant difference in fold change of the number of Dcp1 positive cells (average = 0.7) per anterior lobe (PSC excluded) of col>hdcRNAi (Pcol85-Gal4,UAS-2xEGFP UAS-hdcRNAi) larvae (n = 24) compared to the control (Pcol85-Gal4,UAS-2xEGFP /+) (n = 20). Each dot in the graph represents one anterior lobe. Data were analyzed using two-tailed unpaired student’s t-test, ns: non-significant. (S-T) Silencing hdc in the MZ increases the number of pH3 positive cells (UAS-hdcRNAi/+; domeMESO-Gal4,UAS-2xEGFP/+) (n = 14) (T) in comparison to the control (domeMESO-Gal4,UAS-2xEGFP/+) (n = 12) (S) (red: pH3). Scale bar: 20 μm. (U) A scatter dot plot showing fold change increase in the number of pH3 positive cells (average = 3 folds) per anterior lobe of domeMESO>hdcRNAi (UAS-hdcRNAi/+; domeMESO-Gal4,UAS-2xEGFP/+) larvae compared to the control (domeMESO-Gal4,UAS-2xEGFP/+). Each dot in the graph represents one anterior lobe. Data were analyzed using two-tailed unpaired student’s t-test, *** p ≤ 0.001. (V) A scatter dot plot showing the insignificant difference in fold change of the number of pH3 positive cells (average = 0.8) per anterior lobe (PSC excluded) of col>hdcRNAi (Pcol85-Gal4,UAS-2xEGFP/UAS-hdcRNAi) larvae (n = 20) compared to the control (Pcol85-Gal4,UAS-2xEGFP/+) (n = 26). Each dot in the graph represents one anterior lobe. Data were analyzed using two-tailed unpaired student’s t-test, ns: non-significant.

Fig 6. A graphical summary of the dual role for Hdc in the lymph gland of the Drosophila melanogaster larva.

In the hematopoietic niche, silencing hdc leads to overactivation of the insulin/mTOR pathway. The consequent elevation of ROS levels results in impaired progenitor maintenance and differentiation of lamellocytes. The increase in PSC size as a result of insulin/mTOR activation is compensated by cell death and possibly by G2/M cell cycle arrest caused by hdc loss-of-function. Although hdc silencing also leads to the elevation of ROS levels in the medullary zone of the lymph gland, the insulin/mTOR pathway is not involved in the resulting phenotype. Instead, the JNK pathway, which was previously shown to function downstream to ROS in the MZ, plays a major role. The activation of the JNK pathway downregulates E-cadherin levels thereby damaging cell-cell connections, which leads to lamellocyte differentiation. Although our genetic interaction studies suggest that EGFR functions downstream to Hdc as well, how it is activated by elevated ROS levels, remains to be investigated. In addition to lamellocyte differentiation, silencing hdc in the MZ leads to higher caspase activity and cell proliferation in the lymph gland.