Neuroscience

Prolactin-mediates a lactation-induced suppression of arcuate kisspeptin neuronal activity necessary for lactational infertility in mice

Eleni C.R. Hackwell
Sharon R. Ladyman
Jenny Clarkson
H. James McQuillan
Ulrich Boehm
Allan E. Herbison
Rosemary S.E. Brown
David R. Grattan author has email address

Centre for Neuroendocrinology, School of Biomedical Sciences, University of Otago, Dunedin, NZ
Department of Anatomy, School of Biomedical Sciences, University of Otago, Dunedin, NZ
Maurice Wilkins Centre for Molecular Biodiscovery, Auckland, NZ
Department of Physiology, School of Biomedical Sciences, University of Otago, Dunedin, NZ
Saarland University School of Medicine, Centre for Molecular Signalling (PZMS), Experimental Pharmacology, Homburg, Germany
Department of Physiology, Development and Neuroscience, University of Cambridge, Cambridge, United Kingdom

https://doi.org/10.7554/eLife.94570.1

Open access
Copyright information

Figures and data

Prlr^lox/lox/Camk2a^Cre mice do not undergo the normal period of lactational infertility and the lactation-induced suppression of kisspeptin immunoreactivity is absent.
(A) Kisspeptin immunoreactivity shown in representative photomicrographs from the rostral periventricular region of the third ventricle (RP3V) non-lactating (NL; left) and lactating (right) Prlr^lox/lox control and Prlr^lox/lox/Camk2a^Cre mice (from anteroventral periventricular nucleus (AVPV) region of RP3V). (B) Representative photomicrographs showing mid arcuate nucleus (mARC) of a lactating Prlr^lox/lox mouse (top) and a lactating Prlr^lox/lox/Camk2a^Cre mouse (bottom). (C) Total kisspeptin cell number for the RP3V (NL Prlr^lox/lox (n = 6) versus lactating Prlr^lox/lox control (n = 8) p = 0.0100, NL Prlr^lox/lox/Camk2a^Cre (n = 5) versus lactating Prlr^lox/lox/Camk2a^Cre (n = 8) p = 0.6409). Two-way ANOVA followed by Tukey’s multiple comparisons test. (D) Quantification of kisspeptin fibre density in the arcuate nucleus (Fiji software, measured in percentage voxels per region of interest), showing total kisspeptin fibre density in the arcuate nucleus (lactating Prlr^lox/lox control n = 8, lactating Prlr^lox/lox/Camk2a^Cre n = 7, p = 0.0020, unpaired two-tailed t test). (E) Prlr^lox/lox/Camk2a^Cre mice (blue, n = 8) resume estrous cycles significantly earlier (100% within 6-10 days of lactation) than Prlr^lox/lox controls (grey, n = 10) (p = <0.0001, Log-rank (Mantel-Cox) test). Scale bar image and insert = 50μm. Values are shown as mean ± SEM.

Prolactin action in the brain during lactation is necessary for the suppression of pulsatile LH secretion.
Examples of pulsatile LH levels in the blood from lactating Prlr^lox/lox controls and lactating Prlr^lox/lox/Camk2a^Cre mice that have either been treated with vehicle (sesame oil, s.c., grey, veh) or 4mg/kg mifepristone (in sesame oil, s.c., blue, mif) on the day prior and on the day of blood sampling (2 injections). Asterisks indicate LH pulse peaks as detected by PULSAR Otago analysis. Graphs show LH pulse frequency (B; interaction p = 0.2807, genotype p = 0.0024, state p = 0.8558), and mean LH levels (C; interaction p = 0.8697, genotype p = 0.0031, state p = 0.8586). Lactating vehicle-treated Prlr^lox/lox (n = 8), lactating mifepristone-treated Prlr^lox/lox (n = 8), lactating vehicle-treated Prlr^lox/lox/Camk2a^Cre (n = 8), lactating mifepristone-treated Prlr^lox/lox/Camk2a^Cre (n = 7). Two-way ANOVA followed by Tukey’s multiple comparisons test. Values are shown as mean ± SEM.

Arcuate kisspeptin neuron GCaMP6 population activity throughout different reproductive states in the same mice.
Representative neuronal activity from three Kiss1^Cre mice throughout the virgin, pregnant, lactating, and post-weaning states. The time points monitored in order were: virgin diestrus, day 4 pregnancy, day 14 pregnancy, day 18/19 pregnancy (overnight), day 7 lactation, day 14 lactation, day 18 lactation, 24 hours after weaning (day 22 postpartum), return to normal cycling following weaning (return to estrous cycles), and 10 days post ovariectomy (OVX). Asterisks indicate SEs. Note: dataset from (B) on day 4 of pregnancy onwards and OVX datasets from all mice are on a different y axes scale.

Synchronised Ca²⁺ events are perfectly correlated to pulsatile LH secretion across different reproductive states.
(A) When fibre photometry was paired with serial blood sampling for pulsatile LH secretion, the relationship between SEs and LH pulses was examined. Each of the times an SE was seen during a recording with blood sampling, a pulse of LH was also observed, with 100% correlation (p = <0.0001, Chi-squared test; 73 out of 73 SEs observed lead to an LH pulse). Representative examples of paired photometry and blood sampling are shown from the diestrous state (virgin), from day 7 lactation where no SEs corresponds with no LH release, and from day 14 lactation when SE are beginning to re-emerge. (B) Quantitative analysis of SE frequency per hour across different reproductive states in Kiss1^Cre mice (p = 0.0012, mixed effect analysis (fixed type III) with Tukey’s multiple comparisons tests). (C) Quantitative analysis of SE amplitude of normalised ΔF/F across different reproductive states (p = 0.0118, mixed effect analysis (fixed type III) with Tukey’s multiple comparisons tests). Black asterisks indicate SEs, red asterisks indicate LH pulse peaks as detected by PULSAR Otago analysis. Values shown as mean ± SEM.

Activity of arcuate kisspeptin neurons on day 18/19 of pregnancy
Fibre photometry recordings of mice on the evening of day 18 of pregnancy (0600 hours) to the morning of day 19 of pregnancy day 19 (0800 hours) shows low amplitude SEs. 3-hour section blown up for ease of viewing. (C) No difference is seen between frequency of SEs (per 60 minutes) in the virgin diestrus versus D18/19 pregnancy (p = 0.4063, paired two-tailed t test), however a significant decrease in relative SE amplitude is seen (D; p = 0.0141, paired two-tailed t test). Asterisks indicate SEs. Dotted line in insert of (A) and (B) indicates 3 standard deviations (3SD). Grey shaded region = lights off.

Prlr^lox/lox/Kiss1^Cre mice do not undergo the normal period of lactational infertility and show early reactivation of arcuate kisspeptin neurons prior to estrus in lactation.
(A) Prlr^lox/lox/Kiss1^Cre mice resume estrous cycles significantly earlier (78% within 4-18 days of lactation, n = 32) than Prlr^lox/lox controls (0% by day 18, n = 30) (p = <0.0001, Log-rank (Mantel-Cox) test). (B) Representative fibre photometry traces from day 5 of lactation from either a Kiss1^Cre control mouse or Prlr^lox/lox/Kiss1^Cre mice. Mice with Prlr knocked out of arcuate kisspeptin neurons (Prlr^lox/lox/Kiss1^Cre) show SEs early in lactation, which were not seen until day 14 lactation in Kiss1^Cre control mice. In comparison, the Kiss1^Cre control mouse shows no SEs, as seen in earlier groups sampled on day 7. Asterisks indicate SEs.

Proportion of kisspeptin neurons showing Prlr deletion using RNAscope.
Representative photomicrographs showing RNAscope labelling for Prlr (blue) and Kiss1 (red) in the rostral periventricular region of the third ventricle (RP3V, A) or arcuate nucleus (ARC, C, E), in either intact (Prlr^lox/lox control n = 5, Prlr^lox/lox/Camk2a^Cre, n = 5, A) or ovariectomised (OVX; Prlr^lox/lox control n = 4, Prlr^lox/lox/Camk2a^Cre n = 5, C; Prlr^lox/lox control n = 6, Prlr^lox/lox/Kiss1^Cre n = 7, E) mice. Compared to Prlr^lox/lox control mice, Prlr^lox/lox/Camk2a^Cre mice show a significant decrease in percentage of Kiss1-expressing cells co-expressing Prlr in both the RP3V (B; p = <0.0001) and ARC (D; p = 0.0009) (unpaired two-tailed t tests). (F) A significant decrease in the percent of Kiss1-expressing cells co-expressing Prlr was seen in Prlr^lox/lox/Kiss1^Cre compared to Prlr^l^ox/lox controls (p = <0.0001, unpaired two-tailed t test). (G) No correlation was found between percentage of Kiss1 cells co-expressing with Prlr and the day of estrus return during lactation (p = 0.1912, simple linear regression). A Kiss1-expressing cell was classified as co-expressing Prlr mRNA if the density of Prlr staining was above background. Solid black arrows = doubled labelled cells expressing both Kiss1 and Prlr; black outlined arrows = Kiss1 cells with sparse co-labelling for Prlr. Scale bar = 150μm, insert = 60μm. Values are shown as mean ± SEM.

Mifepristone dose response and effect on litter weight gain.
(A) Mifepristone dose response trial showing dose of 4mg/kg was sufficient to terminate pregnancy in all mice (p = 0.072, Chi-square test, n = 6 both groups). (B) Mifepristone or vehicle injections had no effect on litter weight gain from day 3 to day 5 of lactation (interaction of time x genotype & treatment p = 0.5322; time p = <0.0001; genotype and treatment p = 0.8811; subject p = <0.0001; two-way ANOVA). Values are shown as mean ± SEM.

Gestational and maternal phenotyping of Prlr^lox/lox/Camk2a^Cre and Prlr^lox/lox/Kiss1^Cre mice and their respective Prlr^lox/lox controls.
(A-H) Data show gestational weight gain (A; Prlr^lox/lox n = 11, Prlr^lox/lox/Camk2a^Cre n = 11; E; Prlr^lox/lox n = 31, Prlr^lox/lox/Kiss1^Cre n =27), gestation length (B; Prlr^lox/lox n = 11, Prlr^lox/lox/Camk2a^Cre n = 11; F; Prlr^lox/lox n = 31, Prlr^lox/lox/Kiss1^Cre n =27), number of live pups on day 3 of lactation (C; Prlr^lox/lox n = 11; Prlr^lox/lox/Camk2a^Cre n = 11; G; Prlr^lox/lox n = 31, Prlr^lox/lox/Kiss1^Cre n =27), and litter weight gain between day 3-8 of lactation (D; Prlr^lox/lox n = 8; Prlr^lox/lox/Camk2a^Cre n = 8) or day 3-20 of lactation (H; Prlr^lox/lox n = 22, Prlr^lox/lox/Kiss1^Cre n = 20). There were no differences in any of these parameters (A, C, E, G, unpaired two-tailed t test; B, F, Mann Whitney test; D, H, repeated measures mixed effect analysis, fixed effects (type III) with Šídák’s multiple comparisons test. Grey = Prlr^lox/lox; blue A-D = Prlr^lox/lox/Camk2a^Cre, blue E-H = Prlr^lox/lox/Kiss1^Cre. Some Prlr^lox/lox and Prlr^lox/lox/Kiss1^Cre mice were euthanised prior to day 20 lactation due to COVID-19 lockdown (Prlr^lox/lox n = 5, Prlr^lox/lox/Kiss1^Cre n = 5), due to showing estrus and therefore euthanised 2 hours following blood sampling, or when being used as a control for one of these mice (Prlr^lox/lox n =2, Prlr^lox/lox/Kiss1^Cre n = 2), or due to a litter losing weight (Prlr^lox/lox n =1). Values are shown as mean ± SEM.

Pulsatile LH secretion profiles of Prlr^lox/lox/Camk2a^Cre mice and their controls following vehicle or mifepristone treatment.
Individual LH pulse data from lactating Prlr^lox/lox controls and Prlr^lox/lox/Camk2a^Cre mice treated with either vehicle (sesame oil subcutaneous injection, grey) or 4mg/kg mifepristone (in sesame oil subcutaneous injection, blue) once a day for 2 days prior to blood sampling (2 injections). Lactating vehicle-treated Prlr^lox/lox (n = 8), lactating mifepristone-treated Prlr^lox/lox (n = 8), lactating vehicle-treated Prlr^lox/lox/Camk2a^Cre (n = 8), lactating mifepristone-treated Prlr^lox/lox/Camk2a^Cre(n = 7). Asterisks indicate LH pulse peaks as detected by PULSAR Otago analysis.

Miniature synchronised event-like activity on day 14 of pregnancy does not result in pulsatile LH secretion.
(A) Paired fibre photometry and blood sampling from mouse on day 14 of pregnancy showing miniature SE-like activity have no significant effect on pulsatile LH secretion (red).

Statistics table. Abbreviations for tables below: DF = degrees of freedom; mc = multiple comparison; CI = 95% confidence interval; MW U = Mann Whitney U; MEA = Mixed effect analysis (fixed effects (type III)), RM = repeated measures. Ext = extended data figure.

Statistics table. Abbreviations for tables below: DF = degrees of freedom; mc = multiple comparison; CI = 95% confidence interval; MW U = Mann Whitney U; MEA = Mixed effect analysis (fixed effects (type III)), RM = repeated measures. Ext = extended data figure.

Sign up for email alerts